On History

Abstract

In the late 1970s, students at Essex University wishing to use its DECsystem-10 mainframe for academic work were allocated a project-programmer number (PPN). Mine was [4011,4243]. This came with a fixed set of resources that allowed roughly two hours at a terminal each week.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

CompSoc

In the late 1970s, students at Essex University wishing to use its DECsystem-10 mainframe for academic work were allocated a project-programmer number (PPN). Mine was [4011,4243]. This came with a fixed set of resources that allowed roughly two hours at a terminal each week.

If you wanted to do non-academic work on the computer, that meant joining the Computer Society. Actually, its name might have been the Computing Society; we just called it CompSoc the whole time.

CompSoc made resources available for non-academic use; indeed, if you used your CompSoc PPN for assignments, you would lose it and get zero marks for those assignments. It was solely for play: it could only be used in the evenings and on weekends, when the mainframe wasn’t running to full capacity and so would otherwise have stood partly idle.

Programmers today might have a hard time imagining what this meant for coding. When you have only limited face time on the computer, you have to maximize it. We used to print off listings of our programs and write on them all the changes we were going to make. Because printing also cost resources, sometimes we’d write on printouts that already had writing on them from earlier alterations. We’d make the changes using a text editor, either logged in (which ate up precious time) or via a punched-card batch job that would be run when there was next free capacity. We’d test our programs the same way. As a result, all programmers from that era have a great appreciation for planning out their code, which is not the rather gung-ho, experimental approach that many modern programmers seem to think we had.

Most of the better programmers at Essex University joined CompSoc. I myself signed up because I enjoyed getting computers to do things: I’d learned BASIC at school using a 110-baud modem connected to a computer run by BP as part of a public relations exercise (BP had a chemical works in the area). So it was that once I was at Essex University, I would–through CompSoc–inevitably come to meet other students who were interested in programming for fun.

The Chairman of CompSoc was a second-year student named Nigel Roberts. The secretary was his best friend, Roy Trubshaw.

Did You Know?

Golf was invented in China^{[ 1 ]}, where it was known as Chuiwan (“hitting ball”). There is evidence from the Dongxuan Records ^{[ 2 ]} that it was played as early as the year 945. A Ming dynasty scroll in the National Museum of China^{[ 3 ]}, The Autumn Banquet, depicts a man swinging something having the appearance of a golf club at something having the appearance of a golf ball, with the apparent goal of hitting it into something having the appearance of a golf hole.

Players and Playing

These are a couple of terms I use in this book:

• Player: A person controlling an in-world character.

• Playing: Engaging in an MMO.

I’ve used them for decades, but worry that the colors in them have run so much that it’s hard to tell what they mean any more.

Dr. Toddystone

In my early teens, back in the days when children had spare time, I invented role-playing games.

They’d already been invented by other people, of course—I’m not suggesting I got the jump on Gygax and Arneson—but the point is that I didn’t know about them at the time. It wouldn’t have surprised me to learn of their existence, though—I didn’t think I was doing anything unusual. As far as I was concerned, I was simply using games to tell myself stories.

I had three major (what we’d nowadays call) role-playing campaigns, of which the purest in terms of role-playing (because I didn’t have to pretend to be multiple characters) was called Dr. Toddystone. I was maybe about 12 or 13 when I first played it, hence the embarrassing name; however, in my defense, I wasn’t expecting to tell anyone else about it, except my brother (with whom I discussed game ideas), because it was written for one person and one person only: me.

So, I’ll explain how Dr. Toddystone worked.

Back in the early 1970s, ISO 216 for paper sizes hadn’t officially been adopted in the UK (it had been introduced in 1959, but hadn’t yet achieved critical mass), so I used to make my games using sheets from pads of writing paper that I bought at the local newsagent’s. These measured something like 4½ inches by 6 inches and there were anything between 80 and 140 in a pad, depending on how much profit the manufacturers wanted to make. The pads cost 11p in 1971, immediately after decimalization, but were 14p by the time they stopped selling them in about 1977. I must have bought 50 or more of them in that time.

Anyway, for Dr. Toddystone I took a pad and removed all the sheets. I stuck these together using sticky tape on what was to be the back, as you can’t use felt tips or pencils on sticky tape. I always taped the sheets so the edges were aligned like a grid, as it’s easier to fold everything up that way. The overall shape of my construction was important: sometimes it was long and snaking; sometimes it was bulgy with bits coming off it. The biggest ones (I did this 5 or 6 times in all) were maybe 10 or 12 feet long.

The conceit was that this was an unknown continent, and that Dr. Toddystone was a Victorian-era explorer who was going to explore it.

In laying out the paper, I gave myself ideas of how the geography would go. For example, if I left a sheet out of the middle, that would obviously be a place where I was going to put an inland sea. The paper was only a crude topographical constraint, though: the next step involved actually creating the continent by drawing a map on the paper. This would take hours—days—as I built coastlines, incorporated bays and small islands, river deltas, mangrove swamps, and more. I put in mountain ranges and ran rivers from them to the coast, adding lakes and waterfalls and mosquito-infested marshes. As I did this, my imagination was envisioning the features (even today, writing this, I couldn’t help but add “mosquito-infested” just then). In the process of building up the continent, I got a real feel for the place—its jungles, plains, tundra in the south, mist-shrouded hills, and strange-calling birds in the north.

I populated the continent with various native villages and a few coastal “European” colonies to use as a base. Again, the more I added, the more I understood the place. I knew which tribes were friendly and which weren’t; which were at war and which weren’t; what resources they needed; what strange cultures they exhibited. Some of these promised to be exciting and I hoped Dr. Toddystone would get to encounter them, yet I could never be sure if he would or not.

I won’t pretend that the continents I built were realistic. They made sense to me, but I was only 12 or 13 so I didn’t have all that detailed knowledge of how geographical features actually fit together. In later worlds made for my personal amusement, I’d run some tectonics to get the basic setup and play a game to create a history, but I didn’t do that here. I knew about volcanoes and where they went, so I had those. I didn’t know about volcanic calderas, though, so I didn’t have those.

I should mention that the first time I did this, that’s all I was intending to do: build a continent. It’s fun creating worlds, and so that was my aim. It’s just that, having constructed the world, I was brimming with ideas. It was full of potential; all it took was for someone to go in and release all that penned-in narrative energy. This person was Dr. Toddystone.

Yeah, I’ll tell you about him later.

The Lord of the Rings

Of all the sources that fed into the design of today’s MMOs, the most important is J. R. R. Tolkien’s The Lord of the Rings. The influence of this astonishing work extends not only to fantasy books and all role-playing games, but also to movies and wider culture. Much to the annoyance of literature snobs, it’s consistently voted the UK’s favorite novel, winning the BBC’s huge “The Big Read” survey in 2003 (Jane Austen’s Pride and Prejudice came second)^{[ 4 ]}. The only novel in the English language that has sold more copies is A Tale of Two Cities by Charles Dickens.

It’s my favorite book, too.

The Lord of the Rings has influenced MMO development in three major ways:

• Concept: It showed that the creation of a believable, consistent, self-contained, imaginary world was possible. All children make up their own, pretend worlds. Tolkien demonstrated what could be achieved when adults did the same thing.

• Content: The world that Tolkien describes, Middle Earth, is richly detailed and realized. Generations of MMO designers have plundered it for ideas, whether first-hand from the book itself or second-hand through other Tolkien-influenced works (particularly Dungeons & Dragons). The Lord of the Rings Online uses the content of the book directly.

• Culture: The players of MMOs bring into the virtual world their experiences from the real world. The Lord of the Rings has had such an impact on popular culture that it created the paradigm that is fantasy. There were fantasy stories before Tolkien (Robert E. Howard’s Conan novelettes, for example), but there was no fantasy genre until Tolkien.

The reason I like The Lord of the Rings myself has nothing to do with the story or the writing, which is turgid at times and has Tom Bombadil in it. The only bit of action I really enjoy re-reading is when Éowyn kills the Witch-King of Angmar, and even that’s tainted by Merry’s involvement. No, the reason that I read it three times in my teens and have read it at least once every decade since is because of the first point: the concept. I’d always designed worlds; The Lord of the Rings was the proof that such designs could be beautiful.

Oh, the reason that there’s no spoiler alert here, despite my mentioning Éowyn and the Witch-King of Angmar, is that if you think reading The Lord of the Rings is less important than reading what I’m writing about it, you have really messed-up priorities.

MUD

I’ll be mentioning MUD shortly. Once, it needed no introduction; now, because most MMO players will never have heard of it, it does. Here are the main facts you need to know about it to set the context:

• MUD, an acronym for Multi-User Dungeon, is generally recognized as being the common ancestor of today’s MMOs.

• MUD was written by Roy Trubshaw and Richard Bartle (that’s me, in case you don’t read book covers). Roy began it; I joined him soon after.

• MUD was developed in England. Roy and I are both English. Colour, defence, centre. See?

• MUD was a textual world, not a graphical world. Graphics weren’t widely available back then.

• MUD is now often referred to as MUD1, because MUD was later used to refer to the whole genre. Yes, MUD was a MUD.

Because I co-wrote MUD and have remained working in (what’s now called) the MMO industry ever since, I have some vague name recognition among old-timers. Most players and developers are not old-timers by the definition of “old” you’ll be seeing here; hence, this explanation.

Virtual Whats?

MMOs are virtual worlds. Academics who don’t recklessly invent their own names for them tend to default to “virtual worlds,” and won’t call them MMOs unless one of the following conditions holds:

• They specifically mean game worlds, as opposed to social worlds.

• They are new academics and barely remember Second Life and its ilk.

• They wish to impress non-gamer academics with their grasp of gamer jargon.

• They are clueless.

• They hope to sell a book to people who call them MMOs.

The “world” part of “virtual world” means an environment, the inhabitants of which treat as being self-contained. The “virtual” part means something imaginary that has the form or effect of being real. Together, put more poetically: virtual worlds are places that exist where the real meets the imaginary.

Yes, this does mean they’re not necessarily games. What you do in them can make them games, in which case they’re MMOs. Most virtual worlds were indeed designed to have gameplay and are therefore MMOs, but there are some avowedly non-game virtual worlds: Second Life is probably the best-known example.

Normally, the terminology I use would be virtual world for virtual worlds in general, social worlds for ones along the lines of Second Life with no built-in gameplay, and game-like worlds for the ones along the lines of World of Warcraft that are played as games by most of the people who use them. However, just for you (because I’m using the last bullet from the above list), I’ll generally use MMO in this book rather than “game-like worlds”. Except I might call them virtual worlds anyway, or game worlds when contrasting them with social worlds, or even games if some point I wish to make isn’t MMO-specific.

Oh, and even for social worlds I call the users players. They don’t always like the term, because they don’t feel that they’re “playing”.

Well, they are—because if they’re not, they damned well should be.

High and Low

The architecture of the DECsystem-10 was such that when you ran a program it was split into two parts, a high segment and a low segment, so-called because of which half of memory they each addressed. The former was for program code; the latter for local data.

The high segment was shared. Thus, if 20 people were using, say, the same editor, there would be only one copy of the code in memory (in the high segment) but 20 individual low segments storing each user’s local data (the text being edited, in this example).

Because the high segment was shared, normally you didn’t want anyone writing to it. A change to the executable code for one user would be a change for everyone else running that program, too. You could load data in the high segment, but it was effectively constant because the high segment wasn’t writeable. All users could access it; they just couldn’t change it.

The ability to store shared data in the high segment is what Roy Trubshaw took advantage of when he wrote MUD. Indeed, it’s what initially inspired him to write MUD. He encountered an operation called SETUWP (“set user write protect”) that enabled him make the high segment both shareable and writeable. Instantly, he saw that it could be used to keep a single, shared copy of a virtual world that was the same for every player and could be changed by every player.

It wasn’t the only way to do it: communication through files or through device assignment would also have worked. He was never called upon to think “what can I do with files and device assignment,” though. He was called upon to think “what can I do with SETUWP,” so he did it.

Only?

In 2015, the UK Minister for Culture, Ed Vaizey, told the Develop Conference in Brighton that video games are as important to British culture as cinema.

What? Only as important as cinema?

Dungeons & Dragons

The most venerable role-playing game, progenitor of the whole genre, is Dungeons & Dragons (D&D for short).

It’s hard to imagine today the impact that D&D had on gamers when it came out. It was liberating. It was all about imagination—about sharing a world, about consensus, about living the adventure. People came together to play and could define the boundaries of that play for themselves. They could create their own worlds, their own personalities; they could be the masters of their own destinies; they could make their intelligence count.

The origins of D&D were in miniature wargaming. The way these worked, players would set up armies of hand-painted figures and recreate historical battles such as Waterloo and Gettysburg. The Lake Geneva Tactical Studies Group in Wisconsin developed a set of rules, Chainmail, for mass battles set in the medieval period. One of its authors, Gary Gygax, wrote a 14-page fantasy supplement for it. The fantasy units were like regular ones, but with special abilities—dragons could breathe fire, heroes took four simultaneous hits to kill, wizards could cast fireballs and lightning bolts, and so on. Magical weapons were also present.

Other groups had also created fantasy rules. After all, a dragon in a medieval setting is basically just a World War I bomber with a different name. The crucial addition of a role-playing element came when Dave Arneson took the Chainmail rules and adapted them for his fictional world, called Blackmoor. Instead of controlling armies, players controlled individuals. Gygax saw the potential of this, and he and Arneson collaborated to design D&D. It was first published in 1974 and took the gamer community by storm.

I clubbed together with my brother and a friend and bought my copy on 6^th May, 1976. Yes, I still have the receipt. D&D wasn’t the first role-playing game I’d played, but it was the first I’d played with other people. We played it a lot, too, this being an era when homework wasn’t made of open-ended project work.

The original game came as three booklets: Men and Magic; Monsters and Treasure; Underworld and Wilderness Adventures. These rules were serviceable, but had gaping holes in them; it wasn’t until the Greyhawk supplement arrived that the sweet spot was reached. It had enough extra detail to fill in the gaps and spark ideas, but not so many that it tied the players down.

In 1977, the D&D rules were cleaned up and presented in a more readable fashion as the Dungeons & Dragons Basic Set. We weren’t impressed: they only covered character levels 1 to 3 and seemed to us to be patronizing.

When Advanced Dungeons & Dragons came out in 1978, I was initially enthusiastic. However, the rules for AD&D were far more prescriptive than those of the original D&D, with less scope for improvisation. In my view, the preponderance of rules detracted from the imaginative aspect of play. I still liked it, though, and ran a campaign one summer while I was a student, albeit heavily modifying the system. Several times, we played all through the day, all through the night, and some way through the next day too, all without sleep.

More editions followed, each getting further away from the original. I bought the rule books and read them, but didn’t play them. The second edition of AD&D went with a toned-down concept of fantasy, to address the worries of concerned parents who thought the real deal would turn their children into devil-worshippers. It was less sword-and-sorcery and more of a hotchpotch of myth-and-medieval. More rules were added.

The third edition introduced the d20 concept and was a much less haphazard affair. The rules actually seemed to have been designed with consistency in mind, rather than as an accretion of independent ideas. The fourth edition streamlined the rules, but the result felt more like a tabletop equivalent of MMOs. It polarized player opinion: you either loved it or hated it (most hated it). The fifth edition streamlined the rules still further and went back to a more classic feel. It’s very polished and seems to have reinvigorated the game, which has to be a good thing.

Much has been written about the influence of D&D on computer games, including on MMOs. You have to remember a key point, though: D&D was and remains a magnificent game in its own right. You can’t play a computer game that descends from D&D and expect to understand D&D as a result: you have to play D&D itself. Face-to-face role-playing games, moderated by a person rather than a computer, are limitless. D&D isn’t just some mere waypoint on the road to computer games, it’s a highway to its own destination.

Oh, and it’s fun, too!

ADVENT

Adventure, also known as Colossal Cave and Colossal Cave Adventure, was the 1976 progenitor of the genre of computer games that bears its name: adventures. It was designed and implemented by Will Crowther, based on parts of the longest underground cave system in the world, Mammoth Cave in Kentucky. Fantasy elements and more rooms were added by Don Woods in 1977, leading to the game as it is known today.

At Essex University, because file names were limited to six upper-case characters or numbers, we called it ADVENT.

So, ADVENT was a text adventure. Players typed one- or two-word commands at a prompt that either queried the world state or made changes to it. For example, INVENTORY would list what you were carrying and TAKE LAMP would add the lamp to your inventory. Movement was primarily achieved using compass directions, plus some relational options such as LEAVE HOUSE and limited teleportation. Some commands could be abbreviated, and indeed some had to be: the parser only looked at the first five characters of a word, so NORTHEAST and NORTH became NE and N.

When you played the game, the world from your perspective was described, then you typed something at the prompt. You received feedback from your command, then if the world or your perspective on it changed as a result you got a new description. Here’s a fragment:

YOU ARE IN A 20-FOOT DEPRESSION FLOORED WITH BARE DIRT. SET INTO THE DIRT IS A STRONG STEEL GRATE MOUNTED IN CONCRETE. A DRY STREAMBED LEADS INTO THE DEPRESSION.

THE GRATE IS LOCKED

>OPEN GRATE

THE GRATE IS NOW UNLOCKED.

>DOWN

YOU ARE IN A SMALL CHAMBER BENEATH A 3X3 STEEL GRATE TO THE SURFACE. A LOW CRAWL OVER COBBLES LEADS INWARD TO THE WEST.

THE GRATE STANDS OPEN.

This doesn’t look especially complicated, but if you bear in mind that nothing like it had been written before, you can perhaps appreciate how impressive it seemed to computer users of the era. Programmers were even more impressed, as it was written in FORTRAN.

Many players were inspired by ADVENT to write their own games along similar lines. A decade after ADVENT, text adventures dominated the home computer games market. It was not to last, however: the audience for home computers changed; improved graphics made textual interfaces less attractive; the puzzles in the games became so complicated that they put players off.

Adventure games may now be niche. However, the influence they had on games lives on.

Did You Know?

Golf was invented in France^{[ 5 ]}, where it was known as Palle Mail. Tax records from 1292 show that makers of clubs and balls had to pay a toll to sell their goods to nobles anywhere outside Paris. A c1540 devotional book in the British Library^{[ 6 ]}, Les Heures de la Duchesse de Bourgogne, depicts a man swinging something having the appearance of a golf club at something having the appearance of a golf ball, having the apparent goal of hitting it into something having the appearance of a golf hole.

1978

Back in 2005, the tradeoff between effort and irritation finally tipped: I modified a page on Wikipedia that had been bugging me for ages. Okay, so it was the entry for me, but still.

What I changed was the date of the first MUD. It said 1979; I changed it to 1978. To be fair, the 1979 date was quoting from a Usenet posting that I myself had written in 1990, so you can’t blame anyone for thinking that it was correct. The Usenet posting is wrong, though: it gives the date of the first MUD as Spring 1979 because that was the earliest date I had on a printout of it. Actually, it was more like October 1978.

When I went to Essex University, I signed up to do Mathematics. However, because it had a common first year, I also took Computer Science and Physics. Being a campus university, Essex boasted a thriving collection of student societies, but at the Societies’ Bazaar I somehow missed the one for CompSoc. I kept hearing from other students taking the Computer Science course that CompSoc was worth joining because you got free computer time to do anything non-academic that you wanted, so I went to see the chairman, Nigel Roberts, to sign up. He suggested that we talk while standing in line for the free tickets that the Student Union was giving out for a Lindisfarne Rock Goes to College gig. If we were going to be talking anyway we may as well do it in a queue and kill two birds with one stone.

We got there fairly early (we were within 20 people of the front). Nigel and I are both from the north of England, so it’s natural for us to join queues 45 minutes before the doors open and then complain about the wait. As it was, we also chatted about the kind of things that CompSoc did. I wanted to know if it allowed games, and Nigel said yes, it did. I described to him a transcript of a computer game I’d seen in a postal games magazine called Bellicus. “That’s ADVENT,” said Nigel. “You should speak to Roy about MUD”.

I vaguely recall that Roy did show up briefly to give Nigel his registration card so he could get him a Lindisfarne ticket too, but he was in a hurry to get over to the Open Shop (our one and only computer lab) and would meet us there later.

After we picked up the tickets, we went over to the Open Shop to look for Roy. He wasn’t there—he’d gone back to his flat to pick up some program listings—but a third-year undergraduate friend of his and Nigel’s, Keith Rautenbach, was mooching around. Keith was interested at a technical level in what Roy had done, and had been commenting and making some minor adjustments to Roy’s code. An almost commemorative printout of his final version is the oldest artefact from MUD in existence; it currently resides in the History of Science and Technology Collections at Stanford University, and it’s from this that the 1979 date came (because that’s when I found it on a tape and printed it off).

Anyway, Keith showed me MUD in action, commandeering three or four teletypes (yes, teletypes) to do so. Keith envisaged MUD as a game in which you told characters to do something on your behalf; you didn’t enter the game world yourself, you controlled an ostensibly independent entity (“I am the genie of the watering can”). This was probably due in part to how ADVENT introduced itself. I didn’t see it that way at all, though. I saw it as a way for me to enter a completely new world. Better, I saw it as a way for me to create worlds for other people to enter.

Roy finally showed up, carrying the familiar wad of 11x15-inch green-screen paper under his arm that was the then sign of a programmer. He explained that he’d given Keith the old code to comment and play with because he himself was now working on MUD version 2. Version 1 was merely a proof of concept; he was going to turn it into a full-fledged game. I sought reassurances that it would refer to the player as “you” rather than “I,” and Roy confirmed that of course it would.

So, he asked: did I know anything much about games, then?

1978.

All Programs

At heart, all programs are either a compiler or a database.

– Roy Trubshaw, January 1980

After rewriting the core of MUD:

Or an operating system.

– Roy Trubshaw, April 1980

Criteria

What do I mean when I refer to something as an “MMO”?

Well, I mean a virtual world with gameplay.

Didn’t help, huh? Okay, well I dare say I’ll get around to explaining “gameplay” later, but for the moment, what do I mean by “virtual world”?

Secretly, what I mean by it is anything that descends from the first MUD, but I can hardly say that in public. Besides, I do have a more usable and less self-serving definition.

A virtual world is something with the following characteristics:

• It operates using an underlying automated rule set—its physics.

• Each player represents an individual “in” the virtual world—that player’s character.

• Interaction with the world takes place in real time—if you do something, it happens pretty much when you do it.

• The world is shared—other people can play in the same world at the same time as you.

• The world is persistent—it’s still there when you’re not.

I suppose to be entirely accurate I should add another clause:

• It’s not the real world.

Only if all these criteria are satisfied do you have a virtual world. If even one of them is missing, you don’t. Thus, when I read academic texts explaining that the “first virtual worlds” were cave paintings, or oral folk tales, or peyote-inspired dances, or the I Ching, or theatre, or prayer books, or novels, or telephones, or flight simulators, I sigh.

Let’s put it this way: if you want to call any of those things “virtual worlds,” please supply a new term to describe what objects the six conditions above identify, because whatever objects they do identify, they’re different, they’re special, and they give human beings freedoms and a sense of self that they have never had before in their 30,000-year history unless they were either very rich or were being shot at.

Taking this long view, virtual worlds (and therefore MMOs) are something new and unique. Nothing that has gone before them has delivered what they can deliver. They have so much potential.

I just hope I live long enough to see some of it fulfilled.

Why “World”?

Virtual world is an umbrella term covering both game worlds (MMOs), such as Star Wars: the Old Republic, and social worlds, such as Habbo. Why is the word “world” used there and not, say, “universe” or “space” or “environment”?

Well, as I’m responsible for it, I guess I should explain.

A world in this context is a self-contained environment, the inhabitants of which treat as if it were whole. It’s that notion of self-containment—that the environment is set apart from other environments—that rules out the word “environment” itself.

The word “space” fails because it’s too general: if you were to talk about a “social space” then you could include spaces such as Facebook. Facebook is indeed a space, but it has no connotations of place. It’s the difference between “the world of high finance” and “the space of high finance”: a world is instantiated, whereas a space is where instantiations form.

As for “universe,” well, that would be to imply that there was nothing beyond it. I’m comfortable with calling an MMO set in Colditz Castle a “virtual world,” but a “virtual universe” is overstating its boundaries somewhat. Likewise, if I saw a commercial for a collection of country music pieces, I wouldn’t be surprised if it said “welcome to the world of country music”. I would be surprised if it said “welcome to the universe of country music”. The latter has completely different overtones (and I don’t mean the ones that make your ears throb).

Dinos

People who played textual worlds in the early 1990s were called dinos by later generations (short for “dinosaurs”). Those of us who had been around for 10 or 12 years before the earliest dinos appeared were, of course, amused by this.

Ironically, only people who are themselves now dinosaurs would ever recall the term “dinos” these days.

Versions

As with all software, MUD went through several versions before settling down. It managed four in all, which I’ll describe here not because I want to bore you (although I’m sure I’ll succeed at that), but because later I’ll be mentioning the historical context of MMOs and you may develop a sufficient glimmer of interest to wish to disambiguate between them.

The versions followed (indeed, still follow) the DECsystem-10 naming system. This starts with an integer that indicates the number of times the code has been written from scratch. This is followed by a letter that indicates any major changes that didn’t involve rewrites from scratch but are still rewrites (ports to other operating systems or transliterations into other languages, for example). Following this, in parentheses, is the sub-version integer that indicates the number of releases this version hashad.

So, version 1A(1) was called MUD. It was conceived and written by Roy Trubshaw, with helpful programming suggestions by Nigel Roberts and Keith Rautenbach. Roy wrote it without commenting the code (taking the traditional hacker line that the code is its own comment), but it was later commented by Keith as a tool to help him to understand what it was doing. This explains why its main comment said “MUDD—MULTI-USER GAME OF ADVENTUROUS ENDEAVOUR”—it was Keith making a typing error (the file it was in was MUD.MAC—the .MAC meaning it was written in MACRO-10 assembly language).

Version 1A was not programmed as a playable game: it was a test to see whether the shared memory system that Roy envisaged would work (it did).

Version 2A was also called MUD, and its opening comment expanded the acronym into “Multiple User Dungeon”; the more informal multi- was always used, though, and this became the standard in version 3A. As with version 1A, version 2A was coded in the MACRO-10 assembly language entirely by Roy; its content was run-time programmable by suitably privved (privileged) players. Nigel Roberts and I were privved, and aside from making suggestions actually created some rooms/objects for it. The ox, which is still in the game today, snuck in that way—it was my first MUD creation.

This real-time programmability approach, while common nowadays, used too much computer memory for 1978. Furthermore, as it was written in an assembly language, the code for version 2A itself became increasingly unwieldy. After a year or so, Roy snapped and rewrote the game in BCPL (an ancestor of C) as version 3A. He designed a separate language for defining MUD’s world, which he called MUDDL (“MUD Definition Language”), based loosely on ADVENT’s data format. MUDDL was compiled externally rather than interpreted, then loaded into the MUD engine.

Roy only managed to write perhaps 25% of version 3A’s code (the hardest 25%!) before his final-year exams loomed so close that even he had to pay attention to them, so he handed it over to me to finish. I added the remaining 75% incrementally over the next three years. Fellow CompSoc members Brian Mallett and Ronan Flood each provided some useful hacks to the low-level code that improved its performance (mainly by stopping it from crashing). Although version 3A was begun in 1979, it wasn’t in a playable state until Easter 1980 when I took it over, and therefore this later date is often quoted with regard to the program’s beginnings. I had to recode it as version 3B in 1986 so it could run on the U.S. online service CompuServe, where it was known as British Legends.

MUDDL eventually proved too inexpressive for what I wanted to say, so in 1985 I began work on version 4A using a new MUD definition language that I had designed: MUDDLE. This version of MUD went through four rewrites for various platforms before finally launching as version 4E, which is the one that still runs today. Roy Trubshaw wrote the system tools that accompanied it.

Important! In the 1980s, players of worlds that descended from MUD collectively adopted the name “MUD” as an umbrella term for the entire genre. To avoid confusion, they took to calling MUD itself MUD1. However, MUD1 was not version 1 of MUD: it was version 3. MUD2 is version 4.

Still awake?

Liar!

Exactly When?

I’ve been asked many times for the exact date that the very first version of MUD was written. Yes, it was 1978—but when in 1978?

Okay, so the short answer is that I don’t know the exact date. However, I can make a fair stab at pinning it down.

My first contact with MUD, as I’ve just described, followed my standing in line with Nigel for tickets to a BBC-sponsored Lindisfarne concert. The concert is widely reported as having been broadcast in December 1978, but the date of the recording was 17^th November. We obtained our tickets maybe a month before then.

So, this would only be two or three weeks into term, which at Essex traditionally starts in October. My best guess is that I queued up with Nigel probably in the fourth week, so Roy would have written MUD in the third week. But when in the third week?

All origin stories have inconsistencies. I was once told by another CompSoc member who’d been present when Roy wrote his first (test) version of MUD that there were half a dozen or so hackers sitting around talking code at the time. They’d had this collective idea for writing to shared memory; Roy was merely the first to a keyboard. He got the basic mechanism working fairly quickly with the group’s help, then developed this into version 1 of MUD after about two hours more of programming on his own.

Okay, well I know this story isn’t quite right, because Roy had been trying to get the necessary user privileges to play with inter-process communications for a while. His request had been turned down, so he was actively looking for something that could do the job instead. It’s more likely that the group of hackers was scanning through manuals together and proposing solutions to Roy’s problem, which is why when one of them (Roy himself, I believe) discovered the legendary SETUWP call, he was the one who got to program with it.

Anyway, the thing was, I was given the impression that this happened in the evening, rather than during daytime. I doubt it would have been on a Saturday, because there would have been a Film Society movie on that day and Roy wouldn’t have missed it.

So, taking all this into account, I’d say that the chances are Roy wrote the first version of MUD one evening between Monday 16^th and Friday 20^th October, 1978. Of these, I’d rate the Friday as being the most probable.

The smart thing to do would be to ask Roy, of course, but I did that and he couldn’t remember the date he did it either.

Lines of Code

I once heard that the average number of lines of fully-debugged code the average programmer produces in one day is 60. Blank lines and comments are excluded from the count. Interestingly, the figure of 60 is supposedly independent of programming language.

The code size for the classic version of MUD known as MUD1 was:

• MUDDL code defining the world: 10,347 lines.

• Loader code, to parse MUDDL into usable data: 2,723 lines

• MUD code, to handle I/O and interpret MUDDL data: 9,580 lines

Total: 22,650 lines.

This suggests that MUD had around 377.5 days of work done on it, or just over a man-year.

Actually, though, it didn’t have that much. Both Roy Trubshaw and I are very fast coders and we average over 100 lines of code a day (more when we’re not writing to specification, which was the case for MUD), so at most it would have taken us 7.5 man-months of work to create MUD if we’d been paid to do it.

Elapsed time was longer, of course, because development was spread over a couple of actual years; we had other things to do apart from program MUD.

The First Age—1978 to 1985

For the first seven years of the history of virtual worlds, there was only one: MUD.

Well, actually, no—there were several. MUD was one, yes, but there was also Sceptre of Goth (1978), Avatar (1979), Island of Kesmai (1981), and Habitat (1985).

Those are just the ones that were invented independently during that period; there were others around that were directly inspired by some of them—PIGG (1980), for example, was sparked by MUD. Indeed, the existence of a bunch of are-they-aren’t-they proto-worlds on a system called PLATO (more later) means that, depending on how far you want to stretch your definition of “virtual world,” there could even have been some around before 1978. (I personally don’t think they qualify as virtual worlds, but then I wouldn’t, would I?).

What this tells you is the context of the era. Lots of people had the idea to create what we would later call MMOs, and half a dozen of them successfully built such worlds.

However, only one of these virtual worlds had any meaningful impact on the development of MMOs as MMOs today. The effects of the others endure only as faint echoes, if indeed they endure at all. If we’re talking history and not context, there really was only one virtual world from 1978 to 1985: MUD.

Well, except that Aradath, a 1984 world inspired by Sceptre of Goth, has a tree of direct descendents that includes modern MMOs such as Camelot Unchained. Then again, the influence of MUD’s descendents on Camelot Unchained is probably greater than Aradath’s.

I hope you get the picture that I’m trying to paint here: this is a distant and ambiguous time, and just because a consensus has coalesced to form a clean, accepted canon, that doesn’t mean it’s the whole story.

Nevertheless, whenever the present looks back to the past, it inevitably wants a story, whole or not, and the fact is that MUD really was the progenitor of today’s MMOs. There was nothing like it that its designers had seen before, and there is overwhelming evidence showing how its family tree developed to become the MMOs of today. Any of the virtual worlds I mentioned earlier could have been the one that struck lucky; it happened to be MUD.

We could go back further than 1978, though. I’ve read histories of MMOs that began with prehistoric art, drug-induced shamanistic trances, the Egyptian Book of the Dead, and more besides. It’s only a matter of time before we get one claiming it all started with the possible worlds that came into being during the Planck Epoch (the first 10^-43 seconds after the Big Bang). Yes, they’re part of the causal chain that led to the creation of MUD, but so what?

Less speculative, but still fanciful, histories begin with The Lord of the Rings, ADVENT, Spacewar!, PLATO, and D&D. All of these are indeed important in defining the context, and three (The Lord of the Rings, ADVENT, and D&D) were directly influential in the creation of MUD to greater or lesser degrees. None of them were virtual worlds, though. They’re prehistory, but not history; they may be direct ancestors along other dimensions (for example role-play), but they had no “virtual worldliness” dimension. If you want a history of today’s MMOs in terms of what makes them MMOs (as opposed to anything else), you have to follow their ancestry back until you come to the species line. Continuing further gives you more ancestors, yes, but not ones that are virtual worlds.

So, do that, and the place you’ll stop is MUD.

Yes, I am aware that “stick in the mud” is an appropriate pun here.

I’m also aware that, because I co-wrote MUD, all this is going to look like boastfulness and conceit. It does to me, and I’m the one writing it. To be frank, I find it painfully embarrassing to say all this—so much so that I almost didn’t put anything historical in this book, in order to avoid my having to say it at all. However, history is important, as we shall see time and time again, so I bit the bullet. Neither you nor I have to like it, but the history of MMOs as virtual worlds began with MUD; or, as it’s now known, MUD1.

So that’s why the First Age of MMOs starts in 1978. The reason it ends in 1985 is because, from a revolutionary (rather than evolutionary) perspective, it wasn’t until then that anything else remotely of significance took place.

People played MUD, they liked it, and they figured they could do the same thing only better.

Some could, too.

Great Interfaces of Our Time—Text

Textual worlds describe the environment in words and accept commands in words. Even the most hard-core graphical 3D worlds will still generally have a text box to them for inputting all that tricky freeform speech (although the use of voice, phones, and consoles could eventually change this).

Textual interfaces are the simplest to implement of all MMO-related interfaces, because you can use a general-purpose program such as Telnet as a client.

Here’s an example of a text interface (from MUD):

Road opposite cottage.

You are standing on a badly paved road with a cemetery to the north and the home of a grave-digger to the south. An inscription on the cemetery gates reads, "RESTING PLACE OF LOST SOULS".

*s

Path.

You are standing on a path which leads off a road to the north, to a cottage south of you. To the west and east are separate gardens.

*w

Flower garden.

You are in a well-kept garden. There is an unexpectedly sweet smell here and you notice lots of flowers. To the east across a path there is more garden.

A curious herb gives off a sweet odour nearby.

*g herb

Wolfsbane taken.

*

Text is a very flexible interface, and it benefits greatly from the fact that the form of input is the same as the form of output: words. People think in words, so text is incredibly immersive. As we’ll see later, it has a number of other impressive advantages, too.

Today, however, text is only niche. I don’t need to explain why: nine out of ten of the people reading that transcript from MUD will have already made their mind up never to try it, so they already know the answer.

Happy to read a book, though, eh?

Memories of Memory

So, I went into the attic to try to find my copy of Neuromancer. I didn’t find it, but I did find a box containing the three albums of photographs I took at university while I was an undergraduate.

They’re not wonderful photos. My camera wasn’t good, it had no flash, the prints are decades old and they’ve picked up ink from the photos facing them (when I say “albums,” I mean “scrap books I glued photos into”).

Here’s a typical example: a photo of Roy Trubshaw in the Open Shop (more shortly). The photo is too dark, it’s grainy, its yellows have darkened, and it had bits of scrapbook paper on it until I brushed them off:

My friends in CompSoc thought I was odd for taking photographs of them. Why bother, when we all knew what we looked like? Yet today, students probably take more photos in a week than I took during my entire undergraduate career.

Society back then wasn’t as it is today, because technology back then wasn’t as it is today.

There are 340,690 bytes in that image, or 2,725,520 bits. MUD, the program that Roy and I wrote, was 70k of 36-bit words, or 2,580,480 bits. There are more bits in that image of Roy than there were in MUD’s code.

Those machines behind Roy are teletypes.

Wizzes

In modern MMOs, when you reach the level cap you’re stuck there until the next expansion. In MUD1, when you reached the level cap the game was over.

MUD1 was (well, is, as it’s still extant) a world, though, not a game. The game was over, but play wasn’t. Players who reached the level cap were given vast powers and administrator status. They ran the game for the benefit of the players who hadn’t reached the level cap.

MUD1 levels all had names. You could tell what level a player was because they had the level name appended to their own name—Polly the legend, for example. The level at which points and progress stopped was wizard for male characters and witch for female characters. The collective name for such a character was wiz (meaning witches and wizards).

Later MUDs would call these characters admins or gods. Modern MMOs don’t routinely appoint players to positions of power at all, so their equivalent is the Customer Service Representative, or CSR. CSRs don’t run the world in the same sense that wizzes or admins did, though; they’re more like social workers and police officers than the all-seeing, capricious, supernatural beings of yore.

In MUD1, we also had arch-wizzes, whose relationship to wizzes was roughly equivalent to the wizzes’ relationship to non-wizzes (known as mortals). Arch-wizzes rarely had to do anything, but we had them because if we hadn’t had them we’d have needed them. They started out as being implementers (Roy Trubshaw and I), but later we added other people on the grounds that they knew MUD better than we did.

Did You Know?

Golf was invented in the Netherlands, where it was known as Kolf ^{[ 7 ]}. Because of the danger to pedestrians and windows, in 1480 it was prohibited in the Amsterdam street known as Nes^{[ 8 ]}. A 1755 drawing in the Amsterdam Municipal Archive^{[ 9 ]} by Nicolaas Aartman of the course behind the Stadlander Inn depicts a man swinging something having the appearance of a golf club at something having the appearance of a golf ball, with the apparent goal of hitting it onto something having the appearance of a post.

AAA

In the vernacular, AAA (pronounced “triple A”) means something of the highest quality. It’s the same with computer games, but when the term is not qualified it specifically refers to production quality, not gameplay quality.

A AAA MMO could absolutely suck in terms of gameplay, yet a cheap-and-cheerful one could be brilliant.

ADVENT and MUD

Given that Roy Trubshaw had played ADVENT (he was the first at Essex University to get its notoriously non-obvious last point), it seems fairly clear that ADVENT was a direct ancestor of MUD. It could be argued that merely adding a multi-player dimension to the concept was no different from adding a graphical element to MUDs. In other words, if EverQuest is just a graphical MUD then MUD was just a multi-player ADVENT. After all, the D in MUD stands for “Dungeon” in part because Roy thought that the genre would be called “dungeons” after Zork rather than “adventures” after ADVENT.

This is a reasonable view, although the true relationship is less clear-cut than it might seem at first glance. Roy definitely based MUD’s interface on ADVENT’s, but as all computers back then all had command-line interfaces the idea itself wasn’t new; indeed, it’s hard to envisage how else he might have done it. As for the command set itself, he went with his own but included synonyms for ADVENT’s commands in order that ADVENT’s players wouldn’t have to learn new ones. MUD used GET but ADVENT’s TAKE also worked, for example.

Roy definitely based MUD’s data format loosely on that of ADVENT, so you can certainly contend with some justification that MUD inherited the idea of having a game-specific programming language from it. That has nothing to do with MUD’s virtual worldliness, though. Also, Roy was already familiar with the idea of application-specific languages, having worked with a bespoke, macro-based word-processing program we had at Essex University called FORM. He adapted ADVENT’s format because it best suited his needs, not because he didn’t know about language design and implementation.

ADVENT’s opening instructions began:

SOMEWHERE NEARBY IS COLOSSAL CAVE, WHERE OTHERS HAVE FOUND FORTUNES IN TREASURE AND GOLD, THOUGH IT IS RUMORED THAT SOME WHO ENTER ARE NEVER SEEN AGAIN. MAGIC IS SAID TO WORK IN THE CAVE. I WILL BE YOUR EYES AND HANDS. DIRECT ME WITH COMMANDS OF 1 OR 2 WORDS.

Looking at this, you can see that ADVENT did not take a first-person perspective: you were controlling your representative in its world, you weren’t there yourself. Now although Roy’s initial, test implementation of MUD took the same approach—mainly at the behest of his watching friends—he’d already dropped it by the time I met him. Like me, he wanted players to be in the world, not observers of it.

A key innovation was that MUD was open-ended in a way that ADVENT wasn’t (and adventure games today still aren’t). MUD was about freedom; ADVENT was about puzzle-solving. Although MUD did feature some ADVENT-style puzzles, they were less constrained and more non-linear. This wasn’t just for philosophical reasons, either: in practical terms, if one object (such as ADVENT’s lamp) was the key to advancing, then whoever got it first would lock up the game for everyone else. I suspect that this is why a multi-player version of Zork that was mentioned in the December 1980 issue of Byte magazine came to nothing.

To summarize: ADVENT is undeniably one of several sources that Roy drew on for MUD, but it didn’t give him the idea to write MUD. For me, this marks a clear separation between the two; for you, well, you can make up your own mind.

XP

XP is a shorthand for experience points, or simply points. Tabletop role-playing games tend to call them EPs (which has the benefit of being an actual acronym) or experience.

I put experience points into MUD because I wanted to reward players for their actions in such a way that the accumulation of these rewards would give them goals and commensurate status. I took the idea from wargames and from Dungeons & Dragons (which also took the idea from wargames).

Experience points have been in MMOs ever since. They don’t have to be, but few designers seem to give much thought to that.

Open Shop

Here’s another photo of the Open Shop at Essex University where Roy and I first worked on MUD:

The machines in the foreground are teletypes. They printed onto paper (in capital letters only) at 110 baud. I can type faster than a teletype can show what I’m typing.

Next to the teletypes are wooden boxes. They’re useful as desks, but their main purpose was to hold stocks of paper for the teletypes.

Beyond the teletypes are the punched card machines. You queued, you typed, and you took your cards to the reader. The person manning it fed them in, you waited 15 minutes, and you got your printout. That’s unless there was a Hollerith error because a card didn’t punch properly. You used punched cards because there was no clock ticking on them, unlike the teletypes.

Beyond the punched card machines is a wall with baskets. When you sent something to the lineprinter from a teletype, or submitted a card job for batch processing overnight, it would appear in the basket with your project programmer number on it.

This, boys and girls, is why we had no graphics in MUD1. Our state-of-the-art lab looks as old to you today as today’s computer labs will look 40 years from now.

NOSMOKING was the password to COMPSOC 1, the Computer Society’s master login.

ARPA

When I was an undergraduate, the Internet hadn’t yet been invented. What there was instead was ARPAnet, operated by the U.S. Advance Research Projects Agency. There were only a handful of institutions on it, almost all (unsurprisingly) in the United States.

However, Essex University is located only a 40-minute drive from what at the time was the Post Office Research Centre at Martlesham Heath in Suffolk (it’s now British Telecom’s Adastral Park science campus); the boffins there were working on what they called the Experimental Packet-Switching Service, or EPSS. Because Essex University’s Computer Science department had a long-standing research relationship with them, it was possible for ultra-keen students (such as those in CompSoc) to get access to an EPSS account. Using that, we could hop through a gateway at the University of Kent over to the Stanford Research Institute’s systems. From SRI we could access other universities and research labs using public logins (which was all above board—we weren’t breaking any rules by doing this). I remember visiting BBN, MIT, Stanford proper, Carnegie Mellon, RAND, and UCLA this way.

The CompSoc leadership, Roy Trubshaw and Nigel Roberts, played Zork at MIT and Haunt at Stanford; Zork and Haunt were the only two adventure games around at the time apart from Adventure itself (ADVENT to us), of which we had a local copy.

The Stanford login procedure required us to state our own name, the name of the person we were “hacking for,” and a short description. We didn’t know this beforehand, so when Roy connected to Stanford for the first time (circa 1979), we had to think of answers quickly before the link timed out. I suggested the description of “You haven’t lived until you’ve died in MUD,” which was to become our slogan.

From an historical perspective, it would turn out not to be all that important that we could occasionally get through to overseas computers over what was to become the Internet. What transpired to be the bigger deal was that people from overseas computers could get through to ours.

Sceptre of Goth

Around the same time Roy Trubshaw began work on MUD, Alan Klietz (his surname rhymes with “beats”) independently began work on Sceptre of Goth. Sceptre of Goth wasn’t always its name—it was also known as Milieu, E*M*P*I*R*E and Ghost—but Sceptre of Goth is how it’s remembered today because that was the name its commercial version used. I’ll call it Sceptre for short, as that’s what its players called it.

Klietz developed Sceptre for the Minnesota Educational Computer Consortium (MECC) mainframe. Unlike MUD, his game was fundamentally inspired by both ADVENT and Dungeons & Dragons, as he says himself:

I wondered how to make ADVENT multiplayer. How would interaction work? There needed to be some sort of rule system, one that was simple enough to program into the computers of the day. A student named Mike Pritchard wrote ... a ‘Talk’ program ... and it all clicked. I could write a ‘super Talk’ too, have it show prompts like ADVENT, but be multiplayer using the AD&D rule-set. ^[1]

Because of this, Sceptre played far more like the much later DikuMUD than it did MUD. It became the most popular game on the MECC system—so much so that Klietz had to shut it down between midnight and 5am following complaints from the parents of some of its players. Note that these same hours were roughly the only ones that MUD was allowed to be played by people off-campus, at least on weekdays.

In perhaps the earliest example of what would later be called MUDflation, Sceptre characters obtained gold pieces faster than they could spend them. Klietz decided to reset the database, and built a fiction around why this happened: high-level players would face a series of difficult challenges to obtain the Sceptre of Goth; success would invoke the database-resetting cataclysm. It proved to be a popular idea, too.

When the MECC system was closed down in 1983, Klietz formed a company, Ga¯mBit, with Bob Alberti and two others to commercialize Sceptre. He rewrote the game in C (the original had been in Multi-Pascal) to run on an 8088 (that’s a CPU chip, youngsters). By 1985, Ga¯mBit was grossing $70,000 a year—comparable to what the UK’s top MUDs were making.

Klietz had always been careful with his code. There were no anti-piracy laws at the time, and he didn’t want people ripping off his work (which, his being first class programmer, was frighteningly good). Unfortunately, his fears were soon to be realized.

Maintaining and updating Sceptre was a lot of work, so Ga¯mBit made some hires. The first programmer, hired in late 1984, copied all the system’s software and attempted to sell it through Ga¯mBit’s own chat room. When caught red-handed and fired in early 1985, he immediately used the stolen software to set up a competing system at the rate of $10/month (in contrast with Ga¯mBit’s approximately $2/hour rate.)^{[ 10 ]}

Players flocked to the cheaper system, leading to a collapse of the original’s subscriber base. Ga¯mBit hit back by franchising Sceptre. Soon, there were Sceptre operations in 13 cities (at $10,000/year each). The company was a success, but its principals didn’t really enjoy running a business. They accepted a buy-out offer from a Virginian company called InterPlay (not to be confused with the venerable Californian developer, Interplay Entertainment).

Unfortunately, InterPlay subsequently collapsed, taking Sceptre with it. Had Sceptre not fallen victim to this fate, the course of MMO history would surely have been changed in a profound way. As it was, Sceptre was stopped in its tracks and today’s MMOs are almost all direct descendents of MUD instead.

Note that word “almost”.

Choosing a Genre

If you’re an indie MMO developer and can’t afford to sell an MMO on the strength of its budget alone, what genre do you go for?

Well, you need something that:

• Lots of people like.

• Isn’t available anywhere else at the moment.

• Doesn’t have fans that will be avidly playing something else already.

It’s a set of possibilities that’s getting smaller.

2011

The mere existence of CompSoc meant that when it came to people who programmed for the sheer joy of it, Essex University offered a much more supportive environment than did most other UK universities.

Actually, that’s not all that impressive when you realize that most other UK universities back then didn’t have a computer unless they were rich or needed one for research.

The Computing Service, which operated the DEC-10, was of the collective opinion that people should be encouraged to play with computers if they wanted to do so, because then they would learn more. They therefore allocated CompSoc off-peak resources to use—despite the outrage of the majority of academic departments, which regarded the move a waste of precious computing power. A number of times, some of these departments tried to get the concept of off-peak usage scrapped, so that that they could run SPSS (“Statistical Package for the Social Sciences”) data-crunching batch jobs overnight and have them finish a few seconds earlier the next morning. Others were just annoyed because CompSoc had more resources available to it than they did; it wasn’t that the History Department had much use for computers—it was the principle! However, the Computing Service, especially its head, Charles Bowman, resisted all these attempts to prevent us from having fun; for this, I shall always be grateful.

Ironically, in 2011, I was present at a staff meeting in the School of Computer Science and Electronic Engineering, during which a call was made that the School should reserve its computer labs for people working on assignments only. Encouragingly, I wasn’t the only old-timer to jump down the throat of the person who was fool enough to make the suggestion.

2011 was the project number shared by all CompSoc users. COMPSOC 1, the master account, was [2011,2011] (pronounced “two oh one one squared”).

Okay, so that 2011 was in octal, but still...

PIGG

Having seen our work with MUD, one of my fellow students at Essex University, Stephen Murrell, was inspired to write PIGG (“because it’s a pig to program”).

I mention this not because Stephen Murrell is a bona fide genius (although he is—he was in the Guinness Book of Records at the time for getting the most grade As at A-level in one examination period), nor because PIGG inspired other people to write their own games (it didn’t). I mention it because it used a different method for creating a shared world.

In MUD, multiple people ran separate copies of the same program. This program maintained the virtual world in writeable, shared memory. In PIGG, people assigned their input device (teletype or, later, VDU—“Visual Display Unit”) to one program, which then controlled it along with those of all the other players.

The difference between these two approaches is that in a shared memory system like MUD’s, the virtual world is updated asynchronously. Any copy of the program running can access the shared memory, so there has to be a system of locks in place for when changes are being made. If there’s only one copy of the program running, as with PIGG, you don’t have to lock parts of memory because there’s no competition to access it anyway.

Modern MMOs work in a similar manner to PIGG: the client software connects to a server, which then treats it as an input/output device. There’s no need for any fancy code to make sure that only one command is accessing the world model at once, because there’s only one copy of the program accessing it anyway.

PIGG’s device-assigning idea wasn’t new to members of Essex University’s student Computer Society. Why, then, had no one thought of using it to create a virtual world years earlier? It was only when Roy Trubshaw discovered the writeable, shared memory technology that we got a virtual world.

Well, the solution might have been in plain sight the whole time, but the problem it solved was not. Discovering the writeable, shared memory technique is itself what sparked Roy to make a writeable, shared world.

Island of Kesmai

Another of the (what we’d now call) MMOs created independently of MUD was Island of Kesmai, by Kelton Flinn and John Taylor at the University of Virginia.

IoK was not the first multiplayer game the pair had written. Before it came:

• Air, an aerial warfare game featuring 3D “graphics” rendered in ASCII characters. This was the foundation upon which Kesmai (the name they gave their company) later built its seminal multiplayer graphical dogfight game, Air Warrior.

• S, an eight-player space colonization game. It would become the basis for Kesmai’s hit, MegaWars III.

• Dungeons of Kesmai, a combat-oriented text adventure inspired by ADVENT and Zork.

It could be argued that Air qualifies as an MMO, but it didn’t really have the real-time component nailed. Likewise, S didn’t have the playing-as-an-individual-in-the-game-world element. The most interesting case is Dungeons of Kesmai. It’s clear from Air and S that Kelton and John could do multiplayer, and clear from Dungeons of Kesmai that they could do text adventures. Why, for their next project, did they not combine the two and invent MUDs?

Well, for input they did use an ADVENT-like command-line interface—better, in fact, because you could do things such as OPEN BALM AND DRINK IT. For output, though, they went with the ASCII graphics they’d toyed with in the other two games. Eschewing the 3D look, which was not a success, IoK adopted S’s tessellation approach: the world was made up of squares, each of which was represented on screen by two fixed-width ASCII characters (∼∼ meant water, for example). Thus, although it didn’t use graphics, you could probably call IoK a graphical game; it’s not a huge jump to put a square of water texture on the screen instead of ∼∼. That wouldn’t have helped back when IoK was launched in 1981, though, as most computers could only display fixed-width ASCII characters and no bitmaps anyway.

Although histories tend to discuss IoK only in terms of its pre-graphics graphics, this is somewhat unfair. IoK was an excellent game in its own right. It ran on CompuServe, alongside MUD, and there was a significant player overlap between the two. Although it clearly drew much more on Dungeons & Dragons than did MUD (it had D&D’s character attributes and alignment system), it also had plenty of differences.

For example, its fights were fairly tactical affairs. Movement speed was important, as was positioning. You could be stunned, you could be poisoned, you could fumble your weapon (and have to re-equip it before you could use it again), and your hits could be blocked by armor. Spells had to be “warmed” before casting. This entailed typing a unique incantation (for example, asak nungi irrga luubluyi), unfortunately giving an advantage to players who had communications software with a hotkey capability.

IoK engendered an air of realism: things that “should be” different in small ways were indeed different, and when you tried something because it “ought to” work, it probably would work. Okay, so no one drinks balm in real life (they rub it on), and something like TAKE BERRIES FROM SACK AND EAT IT reads as if you want to eat the sack, but the world itself aimed for verisimilitude. For example, person-to-person communication over arbitrary distances was missing because you can’t do that in real life (unlike in MUD, which did have it).

Bearing this in mind, it should not come as a surprise to learn that in IoK your character could die permanently. This didn’t happen right away, but each time your character was killed there was a chance that some permanent damage to its stats would be done. Once your constitution fell below the minimum, that character was not coming back.

IoK featured many staples of today’s MMOs, such as a bank, shops that buy anything off you, recall rings and character classes. Then again, other D&D-inspired games did too. By the end of the 1980s, IoK was fairly average in that regard; however, it was tuned to perfection and complex enough to merit a 160-page manual.

Oh, there’s one final nice touch I’d like to mention. The players of MMOs have always regarded their world’s designers as the gods of the game’s reality. In IoK they were called ghods, so as not to offend religious people and to spare Kelton and John embarrassment. I always thought that this was rather classy of them—and so much better than MUD’s wizzes.

What Isn’t a Virtual World?

I said earlier that a virtual world has the following properties:

• The world has an underlying, automated physics.

• Players represent virtual characters “in” the world.

• Interaction with the world takes place in real time.

• The world is shared.

• The world is, to at least some degree, persistent.

• The world isn’t the real world.

What this means:

• The world is computer-moderated. Dungeons & Dragons is not a virtual world.

• Players identify with and control a single in-world character when playing. Pillars of Eternity is not a virtual world.

• When you do something, you get a response in a time commensurate with what you’re doing. You issue a movement command, and you move. Planetarion is not a virtual world.

• You and your fellow players are in the same virtual world, not in separate copies of it. What one does, the other sees. The Sims is not a virtual world.

• The world continues to exist in your absence. When the server is rebooted, some things carry over from before the reboot. Counter-Strike is not a virtual world.

• The town or city you live in, which satisfies all the above criteria, is not a virtual world. It’s part of the real world.

What this doesn’t mean:

• Players have to be able to build. They don’t. Killing a monster changes the world; you don’t have to be able to construct a house.

• Players can’t have more than one character. They can. They can even play them separately on two computers if they like. Just because the server doesn’t know you’re not two people, that doesn’t stop it from being a virtual world. The point is, you’re not a party of characters or some strings-pulling remote actor. As for interaction, well yes, you can use VoIP to speak and you can debate on a web site forum, but while in the virtual world it’s all done modulo your character.

• Having multiple servers means you don’t have a virtual world. What it means is that you have several running copies of a virtual world, but each one of those copies is itself a distinct virtual world in its own right.

• If you kill a creature, it’s dead forever. Nah, they can respawn.

MMOs are virtual worlds with a game element built into them. A lot of what’s said about MMOs applies to virtual worlds in general, but not all of it. Some virtual worlds aren’t games, they’re just social spaces.

Exercise: Why isn’t League of Legends a virtual world?

Did You Know?

Golf was invented in Greece^{[ 11 ]}, where it was known as Keritizin. A marble relief^{[ 12 ]}, dated around 510BC-500BC, found on the base of a funerary kouros embedded in the Themistokleian wall of the Kerameikos (the cemetery of ancient Athens), depicts a man swinging something having the appearance of a golf club at something having the appearance of a golf ball, with the apparent goal of ... well, of getting it off some other man doing the same thing. This relief has also been used to suggest that Hockey was invented in Greece^{[ 13 ]}.

Dr. Toddystone

The character of Dr. Toddystone was, of course, based on Dr. Livingstone. I’d played no games that involved the exploration of an unknown (to the explorer—obviously the natives knew about their part of it) continent, but I’d read about Dr. Livingstone and the opening up of Africa. I wasn’t especially a fan of Livingstone’s, but as I’d just built something that was a bit like Africa and was now looking for a means to explore it, well, it was the obvious mechanism to do so. I’d spent so long making such a huge map, had created so many possibilities for exciting things, and had given myself so many ways they could unfold, that I wanted to see what was going to happen there! Dr. Toddystone, as an outsider exploring the continent, was my way to visit the world I had created.

I made myself a compass. No, not that kind of compass, this kind of compass: I took two pins, separated the points by 30mm, and used sticky tape to hold the other ends together. This compass represented how far Dr. Toddystone could travel in a single day, with 1mm representing 1 mile (how’s that for mixing units?). Given that Dr. Toddystone travelled on foot with bearers and pack mules, 30 miles per day is probably a bit too optimistic, but, as I said, I was 12 or 13. I knew how long it took me to walk a mile (15–20 minutes) and worked out Dr. Toddystone’s range from that, factoring in some time for sleeping. Also, it was the absolute maximum distance, so if the terrain was rough he wouldn’t actually go that far (unless I had something special going on, such as a trip down a river by canoe). Sometimes, if he was in a friendly native village or a cave, he wouldn’t move at all.

Finally, I could play. Here’s what I did.

I wrote a diary. Dr. Toddystone began in some outpost of empire and headed off inland. Maybe he was aiming for that mountain in the distance, or perhaps he was following the river, or he could just have been attempting to find where those elephants that occasionally wandered nearby were coming from (he was a scientist, after all). Whatever, he wrote his thoughts down in his diary and then he moved. At the end of the day, he recounted his adventures, then the next morning he’d move again; and so it continued. I recorded his route on the map, wiggling it around if he got lost, making it straight if it were plain sailing or he was attempting to escape from pursuers. He recorded his discoveries, noted his concerns and suspicions, made plans, sought out supplies, followed legends—it was stirring, adventurous stuff! I was experiencing all of it myself, using Dr. Toddystone as my conduit. He wasn’t me—he did things I would never have done—but he was me, in that I was visiting my own, constructed world through being him. It was straight down the line role-playing.

The way the story came out was not always how I had intended or hoped; this is why I thought of it first and foremost it as a game, rather than as anything else. There may have been some particular area where ancient forests grew and where natives spoke in hushed tones of strange, primitive life forms, which I desperately wanted Dr. Toddystone to visit, but he couldn’t because some of his bearers had gone down with a fever and he’d had to head off in a different direction to get the berries he needed to cure them; or perhaps he’d inadvertently offended members of the tribe that lived nearby and couldn’t risk their wrath. There were so many things that could happen, yes, but they interacted. That’s what made it so exciting: I knew what could happen, but not what would happen. The world had a life of its own. I could have cheated, yes, of course, but who cheats at patience?

As soon as I finished the first game, which took me several days (he went from one side of the continent to the other), I started another one on the same map. I did perhaps three or four journeys on that; I could probably have continued with it for another few, but by then I was itching to create a new continent, so I did. I think I undertook fewer discrete journeys on that one—two or three—but they were longer.

Did I re-read the diaries, once I’d written them down? Actually, I don’t believe I did. I used them as reference material, but I didn’t read them as a novel. The purpose of the diary was to realize (i.e., make real) what otherwise was only a mass of interacting possibilities. It constituted a story, but I wasn’t doing it so I’d have a story to read: I was doing it so I could live the story.

The Second Age—1985 to 1989

During the First Age of virtual worlds, the concept was independently invented at least five times. For three of these (Sceptre of Goth, Island of Kesmai, and Habitat), the resulting game was exploited commercially; the developers kept their code and practices secret. For the other two (MUD and Avatar—plus a latecomer, Monster, in 1989), the game was made available for free by enlightened academic institutions and old-era hacker-mentality designers. These games’ developers opened up their code and offered help to anyone who asked.

In MUD’s case, anyone could play if they were able to connect to the Essex University mainframe by whatever means. In Avatar’s case, anyone could play if they had a PLATO console and a direct connection to the PLATO system (which, sadly for Avatar, no members of the general public did).

So it was that the vast, vast majority of people who picked up the idea of virtual worlds and ran with it were MUD players. These started to appear in 1985, and so began the Second Age of MMO history.

First out were Neil Newell’s Shades, Ben Laurie’s Gods, and Mike Blandford’s AMP. They were shortly followed by Pip Cordrey et al’s MirrorWorld, Alan Lenton’s Federation II (there was no Federation I, it started at II), and my and Roy Trubshaw’s MUD2. The number of people exposed to the concept of (what were by then being called) MUDs increased, and as a consequence yet more designers-to-be were inspired to write their own. Because these individuals had to code their own programs from scratch (earlier games tending to be customized for particular hardware), they were able to give full vent to their imagination. The result was a wondrous—and as-yet unparalleled—flowering of ideas, as people pushed at the boundaries of what was possible. It was a very exciting period to be an MMO designer.

It was also almost a dead end.

This was all going on in the UK. In the United States, Avatar was trapped in its gilded cage, and the other “seed” worlds (Sceptre, IoK, and Habitat) were commercial ventures with jealously-guarded code and no cross-pollination. The ongoing evolution of virtual worlds was thus entirely a British thing. This would have been fine if Britain was a comms-friendly environment, but it wasn’t. In the United States, people could make local telephone calls effectively for free; in the UK, local calls cost 40p/hour (in 1986, about £1 an hour in today’s money). Long distance calls cost far, far more. People complained about the price of going online not because the games were expensive (most were free), but because the phone calls were expensive. Some people were racking up bills of £1,000 a month—at a time when the average salary was £750 a month.

With this kind of overhead burdening players, British MUDs had major difficulties attracting newbies who could afford to play. As a result, most commercial MUDs struggled and withered away. Indeed, almost all of them are little more than footnotes when it comes to the history of MMOs, as they had barely any influence on the games we see today. There are exceptions (Avalon is a direct ancestor of both Puzzle Pirates and Earth Eternal, for example), but most of them are of scant lasting significance in and of themselves.

This is something of a shame, in both intellectual and practical terms. Intellectually, because of the sheer inventiveness that was lost; practically, because all the key issues were nailed down in this period but not all were carried through to the Third Age—they had to be (re)discovered anew.

The influence of MUD and its progeny on the history of MMOs could easily have hit a wall at this point. Some commercial MUDs did make the jump to the United States and beyond—most notably MUD1 itself and Federation II—but that fact alone would merely have put them in the same position as IoK and Aradath in terms of their potential for future impact. That’s not what happened, though.

Home computer enthusiasts weren’t the only people who wrote their own MUDs from scratch: so did students, using university machines. Some of these students were aware of, and part of, the wider MUD programming community. Some of them were at sufficiently progressive institutions that they were able to make their games available to the public and other students for free. Some of them wrote portable code that would run on different hardware. Some of them wrote portable code which they gave away.

One of the worlds so created was AberMUD.

MUDTXT.MAS

Here’s a photograph of a printout of MUD1 dated June 12^th, 1985. I think it’s probably the oldest copy in my possession; as I alluded to earlier, I gave my oldest ones to Stanford University Library (they were the first to ask).

The .MAS extension is because it contains a bunch of files combined into one. A program called SUBFIL on the Essex University DEC-10 was used to collect files and create one big file with the sub-files in it; the same program could also split them apart. It’s a bit like a ZIP file, except without the encryption (because if it were encrypted, you couldn’t print it out).

As for why we wanted to package files like this, well the main reason was because each individual file came with an overhead of three blocks (each block being half a kilobyte of 36-bit words). If you had lots of files, those blocks added up and you could easily exceed your storage limit. We therefore preferred to keep our files in .MAS format rather than as lots of smaller files. To this day, I still don’t like creating programs that have separate files for each individual class or function or whatever.

Oh, the reason the job name is THIRD is because it contains the MUDDL code (MUD.TXT) that defines the virtual world’s content. Prior to that, I also printed out the MUD engine code itself and the MUDLIB library that interfaced it to the operating system. These were FIRST and SECOND. If I’d called the jobs anything that looked like MUD, the operators would have spotted what I was printing and told me off for wasting paper.

Also oh, “operators” were the university employees who were allowed to touch the computer and its major peripherals. It took three of them per shift to keep the mainframe and its related hardware up and running. I thought I’d better mention it because it’s not a term you hear a lot nowadays.

The Lessons of Dr. Toddystone

From playing Dr. Toddystone, I learned a great deal. I discovered the power of role-playing—how, by pretending to be someone else, you can be yourself. I understood how games turn a space of potential narratives into actual, personal stories (or “histories,” to be strictly accurate); I picked up some knowledge of geography and cultures, and how they interact, from reading library books for research; and I gained some insight into the nature of imperialism.

Above all, though, I learned what game design is about. It’s about building potentials for players to turn into actualities, and it’s about saying something you can’t say any other way.

I could have written a story instead of drawing a map. I wrote stories all the time; it wouldn’t have been out of character. However, I did draw a map. I did draw a map, because what I wanted to say—“there are so many wonderful things that can happen”—I couldn’t say any other way. I could build the world and write a story set in it, or I could build it and play it.

Following Dr. Toddystone, my future creative direction was never in doubt. I was always, always going to create worlds that people could visit.

Avatar

Avatar was an impressive virtual world written in 1979 by Bruce Maggs, Andrew Shapira, and David Sides for the ground-breaking PLATO system operated by the University of Illinois. Here’s a screenshot:

From this, you can see why Avatar lays claim to being the first graphical virtual world (and hence, if your definition of MMO insists on graphics, the first MMO).

Avatar was actually but the latest of a long line of role-playing games on the PLATO system, which included Mines of Moria, Orthanc, Oubliette, and a handful inspired by Dungeons & Dragons—dnd and The Dungeon being foremost among them. The Dungeon is better known as pedit5, the name its executable file was given so that PLATO’s administrators wouldn’t notice it. The earlier games tended to use a top-down perspective; Oubliette (1977) introduced a first-person perspective, as adopted in the Avatar screenshot (yes, that collection of lines in the middle at the top).

Avatar was a complex game, with challenging gameplay—which is just what PLATO users wanted. It was a roaring success, accounting for 600,000 hours of play from its launch until May, 1985 (about 6% of all PLATO usage for the period)^{[ 14 ]}. Even by today’s standards, it’s impressive; to the players of the early 1980s, it was staggeringly so.

It’s worth asking why, if Avatar was based on Oubliette, which itself descended via Mines of Moria from pedit5, Avatar counts as an MMO when the others don’t. Well, the answer is that some of these others might, too—it really depends on your definition. Avatar does by mine, but Oubliette and its ancestors don’t. Your definition may be more accepting.

In a sense, it doesn’t really matter, though: Avatar’s influence on today’s MMOs is barely perceptible. Even the fact that the term “Avatar” is used to describe the visualization of characters in MMOs isn’t due to Avatar, it’s just a lifting from the same source (Hinduism). Avatar and some of its ancestors anticipated what we’d see in MMOs, but didn’t determine anything about them except possibly their graphics (as one of the first commercial games to use a first-person perspective, Wizardry, was directly inspired by Oubliette).

Avatar was ahead of its time, but trapped in the walled garden of PLATO that gifted it its expressiveness. You couldn’t play Avatar without a PLATO terminal. That was then, though; this is now. Today’s computers are easily powerful enough to emulate the PLATO system, and the Cyber1 project does just that. Head on over to http://www.cyber1.org/ , where perhaps you can try it out for yourself.

AberMUD

AberMUD was written in 1987 by Alan Cox, Jim Finnis, Leon Thrane, and Richard Acott; Rich Salz and Brian Preble further enhanced it. Cox was the primary programmer, his having started work on AberMUD after playing MUD. The name comes from where it was written: the University of Wales at Aberystwyth.

AberMUD was rather rough around the edges. Its architecture owed much to Lance, a MUD being written by Leon Thrane, and its content was a mixture of parodies of other MUDs of the era plus TV shows and stand-alone surreal scenes. On the face of it, then AberMUD was nothing special. However...

In 1988 Cox rewrote the game in C to run on Southampton University’s new Unix machine. This made it portable to any other Unix machine—of which there were hundreds (if not thousands) in U.S. universities. As a result, AberMUD was to become one of the key MUDs in the whole family tree.

ROCK, MIST, BLUD, and UNI

MUD1 needed a lot of memory by the standards of the day to run. Essex University’s Computing Service put limits on the amount of memory that individual programs were allowed to use during the busiest period (prime time), and MUD1 came to exceed these limits. This meant you could still play during the evening and at weekends, but not during the day. “During the day” still applied in summer, when the mainframe was largely idle.

To give diehards something to while away their Resource Control Units when most people were back at home, I wrote a smaller database in MUDDL that conformed to the memory constraints. I called this VALLEY. To create it, I had to go through the MUD database and extract into shared files the pieces that I wanted in both. Having done so, I let other students have access to them so they could write their own MUDs using MUDDL and the MUD1 engine. The four that were actually completed were ROCK, MIST, BLUD, and UNI.

ROCK was the first one written. It was based on the Muppet spin-off, Fraggle Rock, and was a proper, self-contained game. It had no particular angle, being programmed by a bunch of CompSoc MUD1 players for fun. It was fairly good, with some puzzles that stretched MUDDL further than I thought it could go (assembling a drill from components, for example: “There is a bit here”).

MIST was a different world altogether. It was very easy to make wiz through scoring points, so a “game” evolved which was basically all politics. MIST became the personal fiefdom of one of the major figures in the Second Age of virtual worlds, Michael Lawrie (LORRY—he would eventually be responsible for introducing AberMUD to the United States). LORRY would regularly wipe the persona file, dewiz people, promote complete newbies, and so on. Its players loved the sheer anarchy of it.

BLUD was for killing. It wasn’t a game where you could expect to keep a character for very long. Even if the other players didn’t get you, the monsters certainly would. BLUD was written by a single undergraduate in the Computer Science department, although I have a vague recollection that he may have had some link to the Electronics department, too.

UNI was written by a later group of CompSoc members. It started out being a straight implementation of the layout of the Computer Science building, but then parody monsters were added to represent members of staff and it gained a following. UNI was the weakest of the MUDDL worlds, though: as with any in-joke, it only really appeals to those who are in on the joke.

As all professional game designers know deep down, basing a game world on a familiar real-world location for no good reason, as UNI did, smacks of laziness. UNI wasn’t the first world coded in MUDDL to do this, though. Roy Trubshaw based the house in MUD1 itself on the house he used to live in. When I put it to him that this was lazy design, his reply was: “Well yes, I’m lazy”.

MUA

AberMUD was, for a while, known as AberMUG. The reason it was known as AberMUG was that I persuaded its author, Alan Cox, that if he didn’t rename it then people would think “MUD” was a generic term rather than a particular world.

The players just kept on calling it AberMUD.

I myself used to use MUA (“Multi-User Adventure”) as the generic term, on the grounds that this is what Roy Trubshaw would have called MUD if he’d known that adventure games would wind up being called adventure games and not dungeon games.

Needless to say, MUA didn’t catch on despite my efforts, and MUD did become the generic term. The virtual world Roy and I wrote was itself renamed MUD1 (by its players, not by us) to avoid confusion.

When there’s a consensus among players what a word means, they’re going to use it regardless of your wish that they wouldn’t.

And the Winner Is...

Text beats graphics as an interface in every major respect except one: immediacy.

That one difference is enough to consign textual worlds to the margins of history.

Adventure ’89

Adventure ’89 was a gathering in the UK of independent MUD designers/developers that took place, unsurprisingly, in 1989. The event was a showplace for ideas, with perhaps 20 worlds on display, all of them different. I still have some of the fliers I picked up at the event:

When I say that the worlds were different, I mean it. These were some of the genres:

• Federation II—space opera

• The Zone—adult (“score to score”)

• Dark City—cyberpunk

• Strat—holiday on the moon

• Trash—“fire-breathing cabbages and inflatable hover-cars”

• Void—magical adult

• Prodigy—ancient Britain

• Empyrion—underwater city

• Spacers—generation spaceship

Even the fantasy worlds weren’t all the same:

• Gods—end game players can create objects using points given by worshippers

• MirrorWorld—rolling resets (see later)

• Avalon—grid-based in places

• Bloodstone—object decomposition (humans made of 260 parts)

• AMP—objects with shape

• Strata—internal currency

• Warlord—highly combat-intensive

These game worlds were as different as WoW and EVE (although Void was a social world, perhaps closer to Second Life). Why can’t we have equally different worlds now? Where’s the variety?

Three Girls

When Alice arrived in Wonderland, her first words were:

Curiouser and curiouser. ^{[ 15 ]}

When Dorothy arrived in Oz, her first words were:

We will go to the Emerald City and ask the Great Oz how to get back to Kansas again. ^{[ 16 ]}

When Wendy arrived in Neverland, her first words were:

I wish I had a pretty house,

The littlest ever seen,

With funny little red walls

And roof of mossy green. ^{[ 17 ]}

All three girls found themselves in strange and wondrous worlds, but each behaved differently when she got there.

Anarchy

If the name of AberMUD’s designer, Alan Cox, rings a bell, it may be because you’re a Linux user. Alan was for many years the maintainer of the main 2.2 branch of the Linux kernel. He’s an outspoken advocate of free software and argues strongly against software patents. So, he’s one of the good guys, then.

In 2013, Alan was awarded an honorary fellowship by the University of Wales. It’s probably because of all his other achievements, but I like to think that AberMUD might have been a factor.

Decus et Tutanem

Back in the late 1970s, there were basically two main networks for software distribution among the kind of academics who wrote computer games: DECUS and PLATO.

The Digital Equipment Corporation, DEC, was a company that made minicomputers and mainframes. These were very popular in Computer Science departments, in part due to their elegant design and in part due to the fact that DEC wasn’t IBM. DECUS was the Digital Equipment Corporation User Society; it was funded by DEC but run by volunteers, and it distributed software via magnetic tape.

PLATO was a more integrated system, running on Control Data Corporation machines. It was way ahead of its time, having vector graphics as standard when almost everyone else still relied on teletypes. Users accessed a large time-sharing system from pretty well identical terminals, meaning it was relatively easy for someone to write a game in one physical location that could be played by people in other physical locations.

In terms of application software, PLATO was a hotbed of innovation. It can claim to be where message boards, paint programs, instant messaging, and touchscreens first made an appearance. It also saw several multiplayer games, the most celebrated of which, Avatar, would today be regarded as an MMO. As I mentioned earlier, depending on how slack you make your definition of “MMO” and the degree to which you believe Wikipedia, some PLATO games prior to Avatar could be called MMOs, too.

DEC users didn’t have the universal graphics terminals that PLATO users had, so few of their games were graphical. PLATO ultimately gave us Freecell, Battlezone, and Microsoft Flight Simulator; DECUS gave us Hunt the Wumpus, Adventure, and Zork.

DECUS also gave us MUDs.

It didn’t do so directly. It might have done if I’d sent them a copy of MUD1, but I didn’t—I was just a student and had no means to put MUD1 onto the DECUS tapes. When someone who could do it submitted a version without asking me, they took a copy I was working on that had debugging print statements in it and wasn’t really playable. The upshot of all this, then, was that if people wanted a local MUD1-like game, they had to write their own.

This was not without significance.

MUD1 and Avatar were written within a year of each other. That being so, how come today’s virtual worlds descend almost universally from MUD1, with none of them owing a jot of their virtual worldliness to Avatar?

There are two primary reasons.

First, PLATO was effectively an island. Its graphical capability was too far in advance of what was available elsewhere for its games to escape into the environment. All the games written for PLATO were stuck on PLATO until the rest of the world caught up with its graphics.

Second, MUD1 generated more evolutionary pressure than Avatar. On PLATO, the only reason to write your own game was if you were a designer with an idea. Although this was also true in general of DECUS games, in MUD1’s case there was an additional problem: limited access. There was only one instantiation of MUD1 (at least initially), and more people wanted to play it than there were resources available to allow them. Rather than sit around complaining about Essex University’s lack of modems, people went off and wrote their own MUD instead. Shades, Gods, and AberMUD appeared this way, along with dozens more.

The resulting cocktail of competition, inspiration, and experiment led to rapid innovation, both in technology and world design. Thus, when the Internet finally became established, it was MUD1’s descendents that were better positioned to propagate, not Avatar’s (because it didn’t really have any). Later, when commercial developers looked for people with experience in creating virtual worlds, they had thousands of MUD1-descended designers to choose from and only trace numbers of Avatar designers.

So it is, as with much else in history, the way things are with MMOs today is largely to do with blind chance.

LOL

LOL originally meant “laugh out loud”; ROFL was “roll on floor, laughing”. I first encountered them on CompuServe in the 1980s. LOL was something you said (well, wrote—you just regarded it as speech) if you thought something was really funny; ROFL was for if you thought it was utterly hilarious. ROFLMAO was for if you wanted everyone to think you were a superficial jerk.

These days, the usage of LOL has changed. No longer does it mean that you think something is actually laugh-out-loud amusing; rather, it’s a way to acknowledge that you recognize someone’s attempt at humor had vague merit to it. It can also be used to flag that what immediately precedes it isn’t meant to be serious, or that it is but please don’t hit me for saying it.

LOL has proven to be real-world extensible, in that some people now actually use it in everyday conversation. Normally, words such as LOL are present in virtual environments to represent an act of communication that can be expressed in real life but not easily in text (MUAHAHAHA, for maniacal laughter, is like this); however, occasionally one of them fills a linguistic gap that’s present in spoken language as well as written language. LOL is such a term—evidenced by the fact that some people write it as “lawl” (North Americans—it would be more like “loll” in my accent).

Personally, I don’t use LOL or ROFL as I consider both terms dangerously new-fangled. I don’t use smileys either, for the same reason.

Yes, this is indeed a form of snobbery.

Alice

Alice’s view of Wonderland was just that: a land of wonder. She went where fancy took her, with no initial goal save that of satisfying her curiosity.

Alice was engaged on a personal journey of self-discovery. She wasn’t so much curious about the wonders around her as curious about herself. She had to find her own direction in life; Wonderland was the metaphorical place where she did that.

1985 Wish List

Back in 1985, I was asked to do some market research prior to setting up a company to develop MUD2.

To this end, I asked a bunch of senior MUD1 players what might influence their choice of which MUD to play. I sorted these out into categories and got a list of 24. Some of these were features; some were forms of recommendation. I then put this list to the player base in general, asking them to rank each one between -5 (hate) and 5 (love). The resulting average scores were then multiplied by 5 for no good reason and rounded to the nearest integer to avoid decimal places.

Here is the result:

• 25 Intelligent mobiles

• 22 Conversing with mobiles

• 21 More magic

• 19 Cheap per hour to play

• 19 Regularly improved

• 15 Messages to pick up later

• 14 Lots of rooms

• 13 Different start locations

• 13 Reputation of author

• 13 Higher access speed

• 12 Cheap registration fee

• 12 Friend’s recommendation

• 11 Understand complex sentences

• 11 Lots of players

• 10 Increased speed of response

• 10 Free sample game

• 9 Long textual descriptions

• 9 Never crashes

• 9 Special offers for cheap time

• 6 International game

• 5 Lots of books/documentation

• 3 What magazines said

• -3 Built-in adverts

• -3 Graphics

Moral: never entirely believe your players when they tell you what they want.

The Third Age—1989 to 1995

The third age is when the ascent of MUDs became unstoppable. However, it’s also the time of a monumental event in design terms, the consequences of which are still with us.

In the First Age, the concept of a virtual world was invented (multiple times). In the Second Age, it was taken up and developed, but because of the prevailing conditions of the time (economic and technical), it didn’t achieve critical mass. In the Third Age, it achieved critical mass—and then exploded.

AberMUD was one of the more game-oriented MUDs around in the late 1980s. It wasn’t particularly avant garde, but as I mentioned earlier, its designer (Alan Cox) made the critical decision to rewrite it in C. C ran under Unix, the main operating system used by American universities. A strong advocate of free software, Cox released the game onto the nascent Internet. It spread like a rash, and within a year was installed on thousands of Unix systems worldwide.

For the vast majority of players, this was the first MUD they’d seen. In the same way that Second Age MUDs had been inspired by First Age MUDs, AberMUD inspired a new (and larger) generation of designer-programmers to write MUDs of their own; thus, the Third Age came into being.

If that were all there was too it, there wouldn’t be much point in calling this period “the Third Age”. As I’ve described it, it’s basically just more of the Second Age; sure, the dam that was holding back virtual worlds burst, but that alone doesn’t really qualify as an inflexion point. Yet the Third Age is probably the most important period in MMO history to date; what happened then fixed how MMOs are now. Without it, even the concept of an MMO (as distinct from virtual worlds in general and social worlds in particular) would not exist.

The inflexion point was the Great Schism.

The Origins of MUDDLE

To recapitulate: MUD1 hard-coded most of the physics of the game in BCPL (a beautiful language that remains my favorite to this day); the world objects and non-general verbs were defined separately in a data file using a language of Roy Trubshaw’s own design called MUDDL. MUDDL was okay to begin with, but it wasn’t expressive enough to handle many of the concepts we wanted. I kept having to extend it in a series of hacks, until eventually it reached its limit. Every time I wanted to do anything more sophisticated, I had had to code it in BCPL rather than MUDDL; this meant that it was forced on all other uses of the MUD1 game engine (ROCK, MIST, etc.), not just my own MUD and VALLEY.

I therefore decided to rewrite MUD from scratch. I had to determine early on how much to code in C (or, as it turned out, Pascal—thanks, British Telecom) and how much to put into a data file. The more I thought about it, the more obvious it became that everything should go in the data file. If I wanted a generic virtual world engine, how could it be otherwise?

Thus, I created MUDDLE. MUDDLE is data for the MUD2 engine, but code insofar as its programmer is concerned. In other words, it’s a domain-specific language. Whether you want to call it a scripting language or a programming language is up to you; I see it as a programming language, and even have a MUDDLE-to-C compiler for it.

It’s not perfect, though. For one thing, it still leaves too much to the run-time system (e.g., command parsing is done in C, which means that if you want to write a virtual world with no command parsing, or a non-English one, you can’t do it just with MUDDLE).

I’d thoroughly recommend writing a domain-specific language for MMO coding, though. If nothing else, it means you have to understand what it is you need to be able to say in your virtual world design, rather than just trying to say it and then finding out you can’t.

Dorothy

Dorothy’s view of Oz was as a place where she didn’t want to be. She followed a predetermined path—the yellow brick road—to get back to her home.

Dorothy was engaged on a personal journey of self-discovery. She wasn’t so much trying to find the Great Oz as to find herself. She had to follow her own path in life; Oz was the metaphorical place where she did that.

Narrow Road Between Lands

When you entered MUD1, this is the first room description that you saw (cut and pasted from the source file):

Narrow road between lands.

You are stood on a narrow road between The Land and whence you came. To the north and south are the small foothills of a pair of majestic mountains, with a large wall running round. To the west the road continues, where in the distance you can see a thatched cottage opposite an ancient cemetery. The way out is to the east, where a shroud of mist covers the secret pass by which you entered The Land.

I had it printed on some sweatshirts in the 1980s. Here’s one of them:

That block of text on the front is strangely compelling; the players who bought one would say that they were often “read” by strangers while on the tube or waiting in a queue.

One player, Sue the Witch, complained that the words “majestic mountains” appeared right across her, well, majestic mountains. As the first-known example of cross-gender trickery, she later turned out to be a guy named Steve; you have to admire his attention to detail.

I lost about 50p on each sweatshirt, yet people still said I was ripping them off.

Pairs Game

Virtual worlds were independently invented at least six times. See if you can match the inventors to the virtual world they designed:

I’ve put one in to start you off...

MicroMUD

There was a version of MUD written for the Commodore 64 that simulated being an online world. It featured other “player characters,” but they were controlled by the computer. They would “log in” while you yourself were playing. To some degree, they could do most of the things real players could do—even communicate—and they had different personalities. Some would attack you, some would try to help you, some would ignore you, and some hated each other.

Unfortunately, in the time it took for MicroMUD to be implemented the market for text games crashed. Nevertheless, the idea of a single-player version of an MMO remains intriguing.

Would you play a MicroWoW?

The Great Schism

In the first two ages, MUDs had been played both by people who saw them as games and by people who saw them as worlds. The two groups lived happily alongside each other and were mutually supportive. Yes, there were philosophical clashes, but in general each group kept the worst excesses of the other in check; virtual worlds were all the better for it.

AberMUD leaned more toward the game side than the social side. There were MUDs around that were more game-oriented, but overall most had a much stronger social component than did AberMUD. So it was that when AberMUD hit the Internet, not all the social players liked what they saw. They liked they concept, just not the execution.

With TinyMUD, they broke ranks. Written by James Aspnes in 1989, TinyMUD was, as its name suggests, a MUD; however, it also drew from another game, Rich Skrenta’s Monster. Monster had gameplay, but it also allowed player-created content. The second version of MUD had featured this (if you remember, I added an ox to it that way), and so had some Second Age MUDs (MirrorWorld founder Pip Cordrey pushed the idea hard). However, AberMUD didn’t have it, therefore the idea didn’t propagate to the Third Age. The reason that virtual worlds such as Second Life have user-created content is because Monster had it, not because MUD or the games on Pip Cordrey’s IOWA system had it.

TinyMUD was, also as its name suggests, not very large. To make anything of it, people had to take the basic code and extend it with content of their own. This introduced one of the defining features of the Third Age, the codebase. Anyone with modest (or in some cases, no) programming skills could take a codebase and tailor it to make their own MUD.

The viability of this idea was confirmed by Lars Pensjö with LPMUD. LPMUD featured a powerful scripting language, LPC, that enabled people to create sophisticated objects and environments on the fly. Designed to be flexible, LPMUD was used for implementing a wide range of very different, bespoke MUDs. It suffered, however, in that social-oriented players found LPC too much of a chore to learn and game-oriented players didn’t like the fiction-breaking that invariably occurred when other players created content. LPMUD was probably the best codebase of the era for designers, but few players are designers.

TinyMUD wasn’t itself all that impressive, but it did inspire another designer to develop two codebases that were impressive. In 1990, Stephen White improved TinyMUD’s functionality to create TinyMUCK, which he rapidly followed with MOO (“MUD, Object-Oriented”). Like LPMUD, MOO incorporated a full-power scripting language; unlike LPMUD, it was presented in a relatively intuitive way. MOO became the default codebase for non-game worlds of the textual era, and was the basis of perhaps the most famous of all Third Age virtual worlds, Pavel Curtis’s LambdaMOO.

Surprisingly, TinyMUCK wasn’t entirely superseded by MOO and developed a strong following of its own. Furthermore, it moved Larry Foard to write TinyMUSH, which became popular among role-players. By the end of 1990, this made a total of three major social-world codebases (MUCKs, MUSHes, and MOOs), plus one flying the flag for traditional mixed game/social worlds (LPMUDs). The social players had successfully cast off their chains.

This, however, meant that the gamers no longer had to concern themselves with the non-gaming elements of MUDs; they could go as hard-core as they liked. So they did.

DikuMUD was written solely to be a better AberMUD. Although formally a codebase, it came with sufficient content to run as a stand-alone game, and was trivial to install: anyone with access to a Unix (or, by extension, Linux) system could be playing it within minutes of obtaining a copy. Most important, though, was its gameplay: it was utterly compelling. All today’s major MMOs use the DikuMUD gameplay paradigm, which has become so ingrained that at times it’s hard to remember that other approaches do exist.

DikuMUD took the gaming MUD community by storm. With it, the gamers successfully cast off their chains. The social players tried to distance themselves from what was in their eyes a heinous development; the gamers didn’t care, so long as they didn’t have crafters in their midst mewling about being ignored.

Social worlds went one way; game worlds went another. This was the Great Schism that rent the concept of virtual worlds in two, and—sadly in my view—persists to this day.

DikuMUD

I mention a lot of historical MMOs in this book, some of which have had a bigger influence on today’s MMOs than others. However, with the possible exception of MUD itself, the one to which current MMO design owes the greatest debt is DikuMUD.

DikuMUD was a textual world written in 1990 by a group of friends (Katja Nyboe, Tom Madsen, Hans Henrik Staerfeldt, Michael Seifert, and Sebastian Hammer) at the Computer Science department at the University of Copenhagen (Datalogisk Institut Københavns Universitet in Danish, hence the name). It was conceived as an improved AberMUD, but whereas other MUDs of its era were going in the direction of providing greater in-world creation facilities, DikuMUD went in the opposite direction: it hard-coded almost everything.

It hard-coded it well, too: it was easy to install, but any moderately competent C programmer could easily customize it. Many did, and as a result a slew of codebases were created with DikuMUD at their root. There were well over a thousand viable DikuMUDs running in the mid-1990s, with several hundred derivatives, too. Eventually, scripting was also introduced, which resulted in a further flurry of content-creation.

The most important thing about DikuMUD from a modern perspective, however, is not its popularity: rather, it’s the gameplay and other design elements that it pioneered or popularized. Most of the developers of early (graphical) MMOs had played a DikuMUD or derivative, and they took what they learned with them into the commercial environment. To this day, the gameplay of MMOs is very strongly reminiscent of DikuMUD’s gameplay, and even some “new” ideas are merely reinventions of what DikuMUD did years earlier.

Here are some of the things you see in modern MMOs for which you can thank (or curse!) DikuMUD or its derivatives (collected by Raph Koster^{[ 18 ]}):

• Characters:

  • Classes

  • Races

  • Pets

  • Quests (once scripting came in)

• Combat:

  • Procs

  • Cooldowns

  • Aggro

  • Stuns

  • Respawn timers

  • Corpse runs

  • The trinity (tank, healer, and DPS)

• Grouping:

  • Battlefields

  • Raids (sort of)

  • Public quests

• World:

  • Zones

  • Banks

  • Auction houses

  • Housing

  • Instancing

Not everything transported from DikuMUD into modern MMOs lasted. Its death penalties included XP and probable gear loss, for example, which EverQuest had but today’s MMOs don’t. Other ideas such as combat stances, character moods, “rent” for logging off, and admin status for people who “won” the game didn’t stand the test of time either.

Some attractive aspects of DikuMUD have yet to appear in modern MMOs. Procedural content generation is much easier in a textual world than in a graphical world, for example, so DikuMUDs are still ahead there. Also, modern MMO design hasn’t exactly stood still, and concepts such as crafting are done better today than they were in DikuMUD. Furthermore, although the overall hack-and-slash feel of modern MMOs is basically still that of DikuMUD, they don’t simply do in graphics what DikuMUD did in text—they do more besides.

Oddly, despite the fact that DikuMUD influenced the design of modern MMOs a great deal, it wasn’t itself particularly innovative. It drew a lot from Dungeons & Dragons and it drew a lot from earlier MUDs (especially AberMUD). Its contemporary, LPMUD, was far more ground-breaking. However, what it did do was bring all the essential features together and package them in such a way that people could pick them up and run with them. One day, we’ll surely have graphical worlds that afford this kind of creative opportunity to players and would-be designers.

One day...

MUDDL to MUDDLE

This is an extract from MUD1, in MUDDL:

feed nanny victuals     destroy second 682 0

feed nanny antidote     destroy second 682 0

feed nanny flower       destroy second 682 0

feed nanny herring      destroy second 682 0

feed nanny nut          destroy second 682 0

feed nanny pen          destroy second 682 0

feed nanny parachute    destroy second 682 0

feed nanny money        destroy second 682 0

feed nanny gem          destroy second 682 0

feed nanny liquid       destroy second 682 0

feed nanny medication   destroy second 682 0

feed nanny book         destroy second 682 0

feed nanny biscuit      destroy second 682 0

feed nanny paper        destroy second 682 0

feed nanny map          destroy second 682 0

feed nanny tome         destroy second 682 0

feed nanny limb         null    null   684 0

feed nanny corpse       null    null   684 0

feed nanny frog         null    null   684 0

feed nanny bird         null    null   684 0

feed nanny birdofprey   null    null   684 0

feed nanny rodents      null    null   684 0

feed nanny bunny        null    null   684 0

feed nanny vermin       null    null   684 0

feed nanny familiar     null    null   684 0

feed nanny serpent      null    null   684 0

feed nanny none         null    null   687 0

Here’s what it would be in MUDDLE:

{ feed goat edible }:

$(   !! "The goat eats it.*N"

     destroy(second)

$)

{ feed goat animal }:

     !? "It's not a carnivore.*N"

{ feed goat }:

     !? "The goat won't eat.*N"

The nanny goat was what finally made me snap and rewrite MUD from scratch as MUD2.

Lost Worlds

Every once in a while, someone proposes that we create an online museum of virtual worlds, so that future generations can wander around and marvel at how we ever thought they were cool.

I recall a proposal for this in the early 1990s for textual worlds—which should surely have been an easier task than it would be for graphical worlds. The idea was that in order to preserve the past, a meta- world would be created, modeled as a park that included segments of interesting earlier virtual worlds as offshoots. As with real-world archaeology, only the structures would be there, not the functionality.

I was asked to send (and did send) some MUD1 rooms for permanent display in this museum. My aim was to encourage others to donate parts of their MUDs, too; I believed (and still believe) it was important that evidence of the tremendous creativity that went on in MUDs should be preserved before such content was lost forever.

Unfortunately, it’s not just game worlds that fade away when people move on. The museum world itself inevitably succumbed to the same fate as most of the worlds it was trying to archive.

Oh, how the gods of irony must smile.

The Flowering

Instead of asking why we don’t have such a variety of MMOs today as we had in Adventure ’89, perhaps the better question to ask is why did we have it back then? Was it always going to happen? What factors enabled it? Perhaps if we understand the history, we might have a better understanding of what conditions would allow it to repeat in the future.

Okay, so it was always going to happen:

• People saw was possible

• Development costs came down

• There was money to be made

• There were no decent engines available so people had to write their own

• Designing virtual worlds is fun

Do these same conditions apply today?

MMO DSL

A domain-specific language is a programming language designed for one kind of application only. It contrasts with a general-purpose language, which can be used for many applications. DSLs are better than GPLs for the purpose they’re designed for, but they’re not a lot of use for anything other than that purpose.

When designing an MMO, you want some things to be expressed in directly-executable code and some things to be expressed in data for that code. As a general rule, if it changes from run to run, it should be data; if it stays the same, it should be code. For most applications this distinction is fine, but sometimes “from run to run” becomes fuzzy. If, for example, you’re writing a program that is only ever going to be used for one MMO, you could happily hard-code the lot and only save state data (such as whether an individual mob is alive or dead). If, on the other hand, you want your program to work for many different MMOs, you’d want to put a lot more into data.

This leads to a recursion. Here’s how:

• 1. You can, if you want, hard-code your entire MMO in C++. This is bad, because it means you need to recompile it if you want to change any data (e.g., how much damage a sword does).

• 2. You separate your implementation into code (written in C++) and data (written in some format of your own devising). Now you can change the data without changing the program. However, anything that involves execution is still hard-coded; this means if you want to add a new command to your virtual world, you have to code it in C++.

• 3. You create a way of describing commands and store it as data. Now, when you want to add a new command, you only need to modify the data, not the code. You don’t need to recompile the program, because you now have a scripting language that is interpreted on the fly. Some important application-specific functions are still defined in C++, though, and you realize that if you want to use your program for more than one MMO then these should be soft-coded as data, too.

• 4. Your data is now used to define all the functionality of the MMO. The only things left in the C++ are either used by all virtual worlds (e.g., input/output) or they support the interpretation of your data. Your data format is effectively a new, stand-alone domain-specific language for writing MMOs. Let’s call it MMODSL.

• 5. You notice that if you want to make any changes to your data, you have to change your MMODSL code. This is bad, because it means you need to reload it every time you change your data. Now, we’re back at Step 1, except with MMODSL in the position that C++ was in.

There are three main ways out of this.

One way is to decide that yes, you are going to put your data in your code, but you don’t care because you do so at the point where your (C++ or MMODSL) code really is only going to be used for one virtual world. I did this with MUDDLE.

Another way is to separate the scalar data from your MMODSL code (at Step 2 of the second iteration) but take it no further. You can do this because the MMODSL you designed is sufficiently expressive that you never need to put commands into a new data format. Although the data may differ between shards of an MMO, the MMODSL code is the same for all of them. This is how LPMUDs do it.

Finally, you can make the code self-modifiable (i.e., interpreted rather than executed directly), but only modify the data used by MMODSL, not the MMODSL code itself. This means you can make changes to the code but you don’t need to recompile it—you just dump it to disc every so often. This is how MOOs do it.

If you have a DSL for MMOs, you won’t code any quicker in it than if you used a GPL instead. However, each line would carry more expressive power, so you’d be able to achieve more in fewer lines. This is why DSLs are so useful: they let you speak in the language of the application’s design.

Four Uses

In textual non-game worlds such as MOOs and Tiny*s, here’s what people tended to do to occupy themselves:

• Make stuff, either for fun or as conversation pieces.

• Politick about resource allocation.

• Create games.

• Indulge in virtual sex.

The only thing that surprised me about the vibrant sex scene in Second Life was that other people were surprised it sprang up (er, so to speak).

Stock MUD Syndrome

The first virtual worlds—the text MUDs of the First and Second Ages—were written from scratch and were entirely different from each other judged by the standards of today. Yes, the Second Age MUDs drew their inspiration and basic concepts from MUD1, but they headed off in a compass-full of different directions.

Some of them shared code, too. For example, most of the MUDs on the IOWA (“Input/Output World of Adventure”) system were written using a definition language called Slate, for which licenses were available (and actually sold!) at £3,000 each. MUD1’s definition language, MUDDL, was also successfully used to write a succession of fully-fledged MUDs (ROCK, MIST, BLUD, and UNI, as I mentioned earlier). Sharing code did not mean that these MUDs were all the same, though; although some of the concepts were fixed (the IOWA worlds all used rolling resets, for example, which is now the standard for MMOs), the results were all strikingly different from each other. This is because although the definition language was made available, the descriptions of the virtual worlds in this language were not made available (because then people would read them to cheat).

Things began to change in the Third Age, with the appearance of codebases. Codebases all incorporated sample world definitions to get potential developers up and running. Some, most notably DikuMUD, came with a full-blown virtual world that was immediately playable. By downloading a codebase and modifying it, many more people could develop MUDs than at any time previously—surely a good thing?

Well, yes, it was a good thing, but it wasn’t without its downside. See, if people have a choice between writing their own games and modding a ready-made game, many of them will take the latter approach because it’s just easier. They don’t acquire a strong understanding of design fundamentals this way. Also, if large numbers of people downloaded a codebase to mod and then never really put a lot of effort into modding it, the result would be a great wodge of virtual worlds that all looked like the sample one that came with the installation.

Back in the 1990s, this is indeed what happened. Of the several thousand DikuMUDs that were up and running, I’d guess that around three quarters were pretty well clones of one another—a bit like the way today’s MMOs have parallel servers all running the same world. The players and administrators are different, but the rest is the same. For a DikuMUD, as soon as you entered and saw that the main town was called Midgaard, you knew what was coming.

The other quarter of DikuMUDs did take things in different directions; many of these worlds remain unmatched by today’s graphical MMOs for their depth, complexity, range, inventiveness, and gameplay. Yes, I do mean unmatched—and by a considerable margin, too. It doesn’t matter, of course, I know. You’re not going to play them no matter how good they are, because they don’t have purdy pictures.

Most DikuMUDs, though, were not like this. Most were clones, plain and simple. They were repetitive, they crowded out the rest, and although the first one you played would be tremendous fun, the second and subsequent ones would be just more of the same. The barrenness of invention was the stuff of despair for players and designers alike.

From this, we got the term Stock MUD Syndrome. This is a play on the psychological phenomenon of the Stockholm Syndrome, in which people who are held captive come to feel they’re on their captives’ side. People who are imprisoned by identical MMOs with identical gameplay come to feel they’re on their captives’ sides, too, and will often defend them against the slightest hint that they may in some small way be, you know, not perfect.

Tools are good. The more tools we have, the more MMOs we’ll have. Sure, if people can pick up a fully-integrated system for next to nothing, the less likely they are to do anything original with it; however, the more people overall will do something with it, and therefore in absolute terms the number of people who do use it to make something dazzling is going to be larger.

On that basis, I guess I’m okay with them.

Genres and Settings

Currently, the top five most popular MMO genres are:

• Fantasy

• Science fiction

• More fantasy

• Fantasy plus some horror

• Even more fantasy

Okay, so those are fairly broad. Superhero worlds and post-apocalypse worlds are both SF, for example, but as you can see, fantasy dominates. Even something like The Secret World has magic in its modern setting.

I once had a consultancy job in which I was asked to produce genre/setting ideas for particular audiences. Here’s the result:

• Legendary

  • Robin Hood

  • Wild West Outlaws

  • 1001 Arabian Nights

• Historical

  • London, 1966

  • Korean war (MASH)

  • Ancient Rome

  • Age of Exploration

  • Renaissance Europe

  • Colonization of North America (from point of view of natives)

  • Casanova’s Venice

• Victorian

  • Age of Imperialism (Flashman)

  • Grand Tours of Europe

  • Victorian vampires

  • Steampunk

• Action

  • Spy fiction (James Bond)

  • Escapist adventure (Indiana Jones)

  • World War I espionage (Riley, Ace of Spies)

• Inaction

  • Lazy Students

• Science Fiction

  • Escape from Jurassic Park

  • Rampaging Japanese Monsters

  • Asimov’s Foundation series

  • 1950s Science fiction (bug-eyed monsters)

• Crime

  • Hardboiled detective (Sam Spade)

  • Gangsters/FBI (1920s Chicago)

  • Gangs (Warriors)

  • Pimps and Prostitutes

• Modern

  • Cruise Ship

  • Millionaire Street

These are just the ones that matched the target groups—there are hundreds of others I or anyone else can think up.

In terms of genre, much more can be done for MMOs than straight fantasy/SF (or modern for social worlds, or cutesy for kids’ worlds). Historical settings in particular look ripe for exploitation.

A change of fiction isn’t just good for the feel of a world, either: it can suggest new gameplay ideas. If you look at areas where there is no existing paradigm, you have to come up with solutions yourself—you have the freedom to escape from the penitentiary of player expectations as to what a world of that genre “should” look like.

It’s not a case of evolution or revolution, here: you can have both.

Hmm, I suspect my argument might be stronger if I gave an example or two.

Pairs Game—the Answers

Here’s the solution to the pairs game I gave you earlier.

• Avatar was designed and written by Bruce Maggs, Andrew Shapira, and David Sides.

• Habitat was designed and written by Randy Farmer and Chip Morningstar.

• Island of Kesmai was designed and written by Kelton Flinn and John Taylor.

• Monster was designed and written by Rich Skrenta.

• MUD was designed and written by Richard Bartle <wave> and Roy Trubshaw.

• Sceptre of Goth was designed and written by Alan Klietz.

If you don’t remember, who will?

Threading

MUD1 was implemented using a mechanism we’d now call threading. Each player ran their own peer process, which communicated with other processes through shared memory. The operating system’s timesharing mechanism switched between processes, which gave the effect of parallel processing; furthermore, as the CPU was dual-processor; it was possible that actual parallel processing could take place, too. Programmatically, Roy Trubshaw and I therefore took the abstract view and treated each process as if it were running on its own, separate processor.

This being the case, we had to have a lot of signal/wait locking code to prevent memory blocks from being accessed simultaneously by two processes. Otherwise, it would have been possible for one process to start to move an object, then another process to come in and move it somewhere else, then the first process to continue moving it as if it were still where it thought it was. That way lie crashes.

This was something of a pain to program. We had to put in the memory locks manually, as MUDDL was too tied to its interpreter for us to have it insert them automatically. We also had a lingering suspicion that having a sequence of locks on parts of memory ran slower overall than if we’d just locked the entire shared memory each time we had a command to execute.

MUD2 switched to a non-threaded model. As a programmer, this allows you to mess about with objects without the worry that someone else might mess about with them at the same time. The result worked out much faster overall, but it did mean that if one player’s command took inordinately long to execute then everyone else’s commands would have to wait and there could be noticeable server lag as a result. In other words, the game itself was faster but the players occasionally perceived it to be slower.

As it happens, MUDDLE doesn’t really care how it’s implemented, so its run-time system could switch to a threaded model if that improved performance. It’s certainly easier to program MMO content without having to think about threading, though.

Genre Example 1

Many retired people have Internet access. They would be attracted to a genre that isn’t violent, relies on brains rather than brawn, and will attract other retired people. Other criteria that might also work are a hankering for an idyllic past and an opulent or rural setting.

So, that’s an Agatha Christie world, then!

Now if you wanted to make an Agatha Christie MMO, one way to do it would simply be to reskin regular MMO features to fit:

• Tank: The interrogator who has to keep suspects distracted using tools such as small talk, flirting, discussion of their interests, and perhaps physical presence.

• DPS: The investigator looking for clues while the suspect is distracted.

• Healer: The person whose frowns and tut-tuts stop the suspect from doing anything socially gauche (such as breaking off conversation with the tank).

• Rogue: The gentleman/woman thief who breaks into houses, desks, safes, cars, and so on, with only charisma as a defense should servants witness the action.

Okay, so that’s not too bad. It’s already slightly different from the norm in that the healer works by containing the enemy rather than by repairing the tank. However, the setting’s fiction naturally suggests some new gameplay possibilities:

• It would be case-based rather than quest-based. This implies some procedural generation of cases, and that many such cases would be instanced.

• The social world of the 1930s as a backdrop could introduce a contextual social world that could become the preferred experience for some people.

• Having it skill-based or reputation-based, rather than level/class-based, would be more natural. Perhaps introducing social XP leading to social levels would work?

• Espionage could be the end game?

By changing the genre to something a little different, new gameplay ideas soon follow. Rather than trying to find a genre that fits the gameplay, designers should create gameplay to fit the genre.

The Fourth Age—1995 to 1997

Of all the Ages of MMOs, the Fourth is the shortest—but for those involved, the most exciting. It’s the period in which virtual worlds proved their commercial viability.

In the days before regular people had access to the Internet, they used “online services,” also known as “information providers”. The biggest of these, CompuServe, had MUD1 and Island of Kesmai, plus a few other multiplayer games. They did not, however, make a big deal of them; then as now, parents feared that games might turn their children into axe murderers.

Other big rivals to CompuServe—Prodigy, Delphi, The Source—took a similar view. However, there were exceptions. Most prominent of these was GEnie, under the visionary leadership of Bill Louden. Its first major virtual world offering was GemStone ][ in 1988, which became GemStone III in late 1989; it was followed in 1990 by Dragon’s Gate and the proven UK classic, Federation II.

GemStone had been written by David Whatley and his company, Simutronics, in 1987. It was an inspired-by descendant of the First Age virtual world, Sceptre of Goth—indeed, following Sceptre’s demise five people who used to work on Sceptre were hired by Simutronics to create content for GemStone ][ (including Scott Hartsman, who later went on to become Technical Director for EQ2-makers SOE, then later Chief Executive Officer of Rift-makers Trion Worlds). The GemStone games were extremely well received, and despite its textual nature GemStone IV is still thriving today.

Dragon’s Gate was designed by Mark Jacobs, who had written the First Age MMO, Aradath. It was programmed by Darrin Hyrup, who previously had been lead programmer for a Simutronics game, Orb Wars. Jacobs had tried to get Aradath onto QuantumLink (which was later to become AOL) but was turned down. Dragon’s Gate was in part his response to this.

So, in order to follow what happened, you need to know something about the way these online services worked.

Basically, they charged users for the time they were connected. Back when computers were monstrous machines housed in rooms the size of a tennis court, if you wanted to use computer time you went to a computer bureau and paid for the time you used. It was natural that the online companies that grew out of this tradition should use the same pricing strategy.

It also made sense to pay the developers of products for online services a royalty. These were invariably miserable (“How do you expect us to run a service when we only keep 92% of the income our customers generate?”—that kind of miserable). Access to such services was expensive ($12.50/hour for CompuServe in 1990), so per-user profit was good; however, the high price deterred many potential users in the first place, so overall income was merely modest.

Then, along came the World Wide Web. Suddenly, lots of people wanted to access the Internet. A price war between the big online services in 1993 meant that now they could afford to, too. They flocked to the service providers, seeking things to do, and found, oh, look—games!

AOL noticed a sudden boost in player numbers for its innovative-but-flawed MMO Neverwinter Nights, and realized what was going on. It quickly lured GemStone III, Dragon’s Gate, and Federation II into the AOL fold, by the simple technique of offering their developers a fair royalty. As the newbie hordes poured through the gates, usage rose spectacularly.

Thousands of players, a good royalty, per-hour charging—these were the ingredients that stoked the fires of the Fourth Age. Even the “failing” Neverwinter Nights, capped at 500 simultaneous users, grossed $5m in 1996.

For the 18 glorious months of the Fourth Age, the developers of AOL’s MMOs wore money hats. Then, it all fell apart. Smaller companies started charging flat rates for Internet access and in late 1996 AOL had to follow suit. They couldn’t keep paying royalties on a per-hour basis, and by mid-1997 the show was over. Nevertheless, GemStone III still managed to gross $10m from 2,500 to 3,000 simultaneous users that year.

The Fourth Age of MMO development may have been short, and it may have only involved a handful of MUDs, but it left one, lasting legacy without which we wouldn’t have the MMOs of today: it caused the regular game industry to notice virtual worlds.

Wendy

Wendy’s view of Neverland was as a place of fantasy—her own. She made up the world herself, instantiating her imagination by telling stories.

Wendy was engaged on a personal journey of self-discovery. She wasn’t so much envisioning Neverland as envisioning herself. She had to find out who she was in life; Neverland was the metaphorical place where she did that.

Great Interfaces of Our Time—Text with Fixed Graphics

The general-purpose comms programs that players used to connect to MUDs didn’t support graphics. However, if you wrote your own client (or front end, as they were known), you could add graphics to it.

Gemstone III took this approach. Here’s a screenshot that appeared in some 1993 GEnie advertising material (yes, I do keep this stuff for decades):

As you can see, there’s an illustration of a forest in the top-right corner; this would change for other types of terrain. The rest of the title bar shows health and mana bars that would be familiar to players of today’s MMOs, plus other useful at-a-glance stats.

The rest of the screen is text. I nearly wrote “just text,” but there was no “just” about it: this was a quality game. Check out Gemstone IV to see what I mean.

Originality

What color are health and mana bars?

In general, health bars are red or green (or both if you’re color-blind) and mana bars are blue. If there’s no mana bar, the health bar will be green but get yellow then red as damage is taken.

Why is this? They don’t have to be colored this way. Neither do they have to fill up left-to-right, nor even be bars in the first place. Aren’t designers capable of originality?

Well yes, of course they are. They deliberately follow the herd, however, because that’s easier on the players. If a user interface item works and the players understand it, you need a very good reason to make them learn a new one.

Originality purely for originality’s sake is not always a great idea. Good designers know this.

PvP v PvE

In MMOs, players sometimes compete against other players and sometimes compete against the environment. The former is PvP (player versus player); the latter is PvE (player versus environment).

Although, strictly speaking, PvP and PvE can cover all manner of activities, combat is the usual differentiator. If you spend your time fighting other player characters, that’s PvP; if you spend your time fighting computer-controlled enemies, that’s PvE. A server that allows PvP combat is called a PvP server; one that doesn’t is called a PvE server.

Some people love PvP and some people hate it. The same can be said of PvE. It’s very hard for designers to balance an MMO for both of them. It’s not easy for designers to balance an MMO at all, come to that.

Client/Server

Most MMOs use a client/serverarchitecture. In this, the players connect their computers (the clients) to another computer (the server). Both these terms (client and server) can refer either to the hardware or software involved. Either way, the key to this architecture is that the server alone makes all the gameplay-critical decisions, leading to what is called a centralized system.

Here’s what one looks like (or did when we used CRT monitors):

The server reflects the true state of the game world; the clients perform input/output for the server, and at best each client can only show what was true when the server last told it anything.

This architecture shares some load, but for a regular MMO the server will have to be many times more powerful than the disparate computers that are being used as clients. Indeed, for most modern virtual worlds the server isn’t itself a single computer but a cluster or (for very large worlds) a grid or cloud. In any event, from the server’s point of view, there’s a huge bandwidth cost in dealing with all those incoming connections—far more than in a distributed system.

In other words, it’s expensive.

Given this, then, why does almost every MMO out there nevertheless use a client/server architecture?

There are some legal and business reasons for it, to do with controlling the code and providing workable customer service, but what about the gameplay reasons?

Well, it’s very hard for players to cheat on a client/server system unless the developers are naïve or foolish enough to allow the client to make authoritative decisions. When you remember that “cheat” here can mean “steal all your players,” that’s an important factor.

This approach is also highly resilient to client-side crashes; yes, if the server goes down then that’s it for everyone, but if my PC dies then it’s only me who’s affected. The server for a well-maintained client/server system can persist indefinitely (we’re already talking decades in some cases).

The fact that client/server is far easier to program than any of the alternatives is merely a happy coincidence.

ADVENT Coda

I mentioned earlier than although ADVENT was an influence on Roy Trubshaw when he created MUD, it wasn’t an inspiration. His inspiration was the coming-together of a number of factors at the nexus that was his discovery of the SETUWP operating system call.

Roy wasn’t the only person who wrote MUD, though. I co-wrote it with him. It’s reasonable, therefore, to ask if I was inspired by ADVENT in a way that Roy wasn’t, as this would add weight to the suggestion that MUD could be regarded as the spiritual child of ADVENT.

When I was at school, well before I came across ADVENT, I wrote a program in BASIC called TALKER that tried to hold a conversation with the user. It essentially worked using keyword recognition; appropriate responses were composed from canned, pre-written fragments. It was hard-coded and naïve, but surprising effective for about a minute or so. I did have a notion that I could make the computer interlocutor act as the gamesmaster of a role-playing game, but could see it would take up more computer time than we were allowed.

When I first came across a page-long transcript of ADVENT I was moderately pleased: it showed that what I’d imagined could be done could indeed be done. That’s why I described the transcript to Nigel Roberts while waiting for free tickets to see a folk/progressive rock band in October 1978, and why he suggested I should speak to Roy Trubshaw about MUD.

So no, it wasn’t an inspiration for me; it was more a proof of concept.

Genre Example 2

Suppose you wanted to create an MMO for the above person.

Okay, so she aches for what once was and hopes for what could once more be; she prefers mind over muscle, the subtle over the crass. The key word for her is intrigue. That says something like the Scarlet Pimpernel era, or Eleanor of Acquitaine, or the Doge of Venice.

In an historical context, gender usually plays a larger part than in SF or fantasy. Assuming we didn’t want to take that particular bull by the horns, we might perhaps consider the revolutionary France setting: we could use the liberté, égalité, fraternité ideology as cover to give female characters on the revolutionary side freedom of action; on the aristocratic side, the privileges of wealth would achieve the same objective. Okay, so let’s go with that, then.

The gameplay could involve either freeing aristocrats or imprisoning them. There might be some fancy swordplay (although never ending in death or injury; people still had standards back then), but there could also be some rather-more-unusual verbal battles.

Verbal battles have the potential to introduce some new techniques to combat. For example, if you say something that embarrasses your opponent, the damage it does is relative to how many people they have on their side. If they came with a lot of friends, the damage to their ego will be much worse than if only you and they heard the remark. Sarcasm works as a good riposte to bombast, but a muted one to earlier sarcasm. A joke will hurt, but only if it’s not been used much before; otherwise, it could backfire and hurt you instead. There’s a lot that could be done here: rudeness, playing dumb, accusations, laughing remarks off, impregnable logic, breaking down in tears; I’m sure you can think of more. Regular MMO combat has a background stream of damage that is then modulated by individual commands performed independently and simultaneously by the combatants. A damage meter is reduced until it reaches zero for one side, whereupon combat is over. The amount by which a meter is reduced is determined by a rotation of the commands available to a player: 3 clicks of this then 1 click of that, then back to 3 of this, maybe 1 of those if it comes off cooldown—all the while not standing in the fire.

What we have here with verbal battles is much the same, except it uses a turn-based approach in which what you respond is based on what your opponent just said. You have to choose from the possible things you could say to counter what they said, while guarding that you’re not going to give them an opening to say something devastating in response.

There are plenty of games that have this kind of system in place (not necessarily for verbal combat), but it’s not exactly common in MMOs. That’s not what I’m getting at here, though. The point is, this came from the choice of genre. Switching from fantasy to France, 1789, naturally led to a different combat mechanic.

Genre change can do that, because it brings with it new fiction that has different requirements to traditional fictions. However, this does mean that you have to uphold the integrity of the fiction rather more than designers tend to do when it comes to fantasy.

Tanks

Yes, you know what the term tank means: it refers to the person who stands there taking all the damage while the healer clears it and the DPS deals it to the opponent.

What you might not know is that originally it referred to the kind of over-powered character you got in some classless systems, who could take damage, heal damage, and deal damage all at the same time. People would argue against classless systems “because then you get tanks”.

Oh, how times and terms change.

Berserkers

Some of the things I added to MUD1 I later took out. Primary among these was the berserker character option.

There was a command you could use to make your character a berserker. As a berserker, you got a special prompt (a > instead of a *) and double points for killing other player characters. However, you couldn’t flee fights and you could never become a wiz (that is, a witch or a wizard—the game’s administrators).

Berserkers were a failed experiment. They were out-and-out killer characters, and they exerted influence way beyond their number. The problem wasn’t so much with the mortals who ran berserkers (they kept to the “no bullying” rules), but with the wizzes who were supposed to come down hard on mortal excesses. See, then as now, wizzes had secondary characters (alts, we’d call them today) and some wizzes made a berserker. When these berserkers misbehaved, wizzes were reluctant to interfere: they knew that the player running them would log back immediately on their wiz and argue about it.

I didn’t really expect berserkers to be a success when I implemented them. It’s not that I deliberately programmed them to fail or anything; rather, I was fairly certain that they would be abused. It turned out I was right. In that case, why did I add them in the first place?

Well, I added them for two reasons: to undermine any notion that D&D-style character classes should be introduced into MUD1; to make clear to wizzes what constituted responsible behavior on their part. I never really needed the former argument, but I often used the second one. Eventually, we wrote a document—the Good Wiz Guide—that formalized how wizzes should use their powers; it’s still in use today.

Interestingly, Sceptre of Goth had a berserker-like “barbarian” class that was eventually restricted for similar reasons: it was over-powered, and when groups of barbarians came together they could (and therefore did) defeat anything that stood against them. In that respect, they were rather like real barbarians.

Great Interfaces of Our Time—Graphics Made of Text

When you’re using a connection that can carry ASCII characters at a rate of about 30 per second, maximum, there isn’t a great deal of scope for graphics. If potential players of your virtual world only have general-purpose client software (and we’re talking raw text terminal protocols here, not even Telnet), then what can you do?

Well, you can use ASCII characters to draw your virtual world.

Island of Kesmai, by Kelton Flinn and John Taylor, did just this. Here’s a screenshot (missing two lines of stats at the bottom that would mean I’d have to shrink everything so small to fit them in that you couldn’t read any of it without the aid of an optical instrument):

[][][][]--[] orc  sword  shield  chain

[]A     S [] 1 A Jennie.c

/ >  ++++ [] 2 A 2 skeletons

[]   ++++ [] 3 B trolls

[]B    dn []

[][][][][][]

Swing hits with moderate damage

Skeleton is slain

Orc is blocked by your armour

Troll: kia ardata luuppatar ne

>throw bottle at troll

As you can see, the display had a map of cells (squares, but IoK called them hexes), each one comprising two fixed-width ASCII characters. [] was a wall, for example. Labels were used to refer to the contents of squares, so we can see that A contains Jennie.c and two skeletons (probably bad news for the skellies, as Jennie.c was one of the top players).

IOK was written in 1981, completely independently of MUD. It was a big hit on CompuServe, even though to play there at 30 characters per second (300 baud) would have cost you $6 per hour in the late 1980s—and more than double that if you used a “fast,” 1200-baud modem.

Designing Is Fun

The main factor that made the Adventure ’89 flowering of creativity inevitable was that designing MMOs is fun. Indeed, if you think about it, it’s the only reason anyone would design them—they’re not something the real world needs. As we’ll see later, though, “fun” can mean different things to different people.

Of course, it isn’t just anyone who’d design and implement an MMO: it takes a special kind of person to do it. First and foremost, the people who attended Adventure ’89 wanted to create worlds. Sure, many wanted to make money, too, but that was basically an excuse: the prospect of gold raining from above was how they rationalized spending their time creating worlds. Really, though, they created worlds simply because for some people creating worlds is fun.

They were those kind of people. Are you?

Genre Example 3

How about an MMO in which the aim was to escape from a prisoner of war camp?

This would attract some of the more explorative and crafting-oriented players who are ill-served by today’s combat-focused MMOs. The bulk of the gameplay would concern sneaking and hiding and distracting, rather than combat, because the guards have guns and you don’t. So, we’re talking a stealth MMO (without zombies, for a change).

Interestingly, when you reach Blighty, it’s game over for that character. There’s no endgame: you can play another character or leave on a high.

Again, see how a simple change of setting brings with it implications from the fiction that affect the gameplay.

There’s no design law that says MMOs have to play exactly the same way.

The Tragedy of Geography

MMOs (in the form of MUDs) were invented in the UK, but we don’t have many major MMO developers here. Why is that?

Well, it’s because no one with money in the UK at the time MUDs made a splash believed in the concept (or, if they did, they made themselves so difficult to find that I never came across them). The underlying reason for this is the financial conservatism of the UK’s banks and other investment organizations, which are famously reluctant to put money into new ventures unless you can prove categorically that you don’t need it. Had MMOs been invented pretty well anywhere else in the developed world, they would have been given a much better reception than they were in England.

The thing is, though, that when they were (re)invented anywhere else, they never became strong enough to last. Almost every virtual world in existence can trace its genealogy directly to MUD1, not to other first-generation worlds. This is in part because of the very forces of financial conservatism that stopped us from commercializing MUDs in the first place: we were free to do whatever we liked with them. In the United States, designers spent time writing worlds they could sell; in the UK, we knew we hadn’t a hope of doing that so we wrote them for fun instead. The art trumped the commerce, at least until the commerce could afford to brush aside the art.

So it is that we conformed to the traditional British stereotype of being great at inventing things but hopeless at exploiting them afterward.

Adding Coffee to Tea

If you add a drop of coffee to a cup of tea, few people will be able to taste it. If you add a couple of drops more, suddenly everyone can taste it. This is fine if you like coffee, but not so fine if you paid to drink tea.

Add too much of the real world to an MMO, and you’ve got the same situation.

Creating Realities

People have always created their own realities. What makes MMOs different is that now, for the first time, other people can enter those realities.

Great Interfaces of Our Time—Tile-Based Graphics

With your own client software and computers that can display graphics, it’s not hard to see how Island of Kesmai could easily have been converted to use images for each square instead of ASCII characters.

Or, if the Kesmai Corporation is being slow about it, you could write your own IOK-like world and beat them to the punch.

Here’s Neverwinter Nights, as it looked in AOL’s advertisements in 1992:

No, this isn’t the NWN released by BioWare in 2002, nor the Neverwinter MMO developed by Cryptic in 2013; this is the Stormfront NWN from 1991 designed by Don Daglow, which I mentioned earlier when describing the Fourth Age of MMOs.

You’ll notice that although the map is actually a flat tessellation of squares, some effort has been expended to give an appearance of perspective. While not entirely convincing, NWN was better at this than most of the other attempts that appeared at around the same time. Here, for example, is Kingdom of Drakkar from some 1993 MPG-NET publicity matter:

The mixture of oblique-projection walls and side-on animals and player-characters is a little, well, unimmersive.

The big filler image at the top of the right panel was for further details. If you examined your inventory, for example, you’d get this kind of thing:

Note that the command line at the bottom still uses text (it was the top line for NWN), and there’s a separate panel for incoming messages. Even though KoD appeared only a year or so later than NWN, its interface is a lot more detailed and fussy—showing how quickly things began to move once home computers were capable of displaying half-decent artwork.

There is a modern take on this kind of thing with voxel-based worlds such as Minecraft. These use cubes instead of squares but otherwise work on similar principles except that no one in them looks if they’re lying on their back unless they are, in fact, lying on their back.

Genre X

Suppose you’re fabulously wealthy and decide to create an MMO. What genre would you choose?

Let’s say you choose genre X. Of all the genres you could have gone for, you went for X. You don’t need the money (you’re fabulously wealthy), so why did you choose X? Well, X must enable you to say something that you can’t say as well (or perhaps at all) in any other genre. Otherwise, you’d have chosen Y.

Designing a virtual world is an art: there are thousands of decisions you have to make, some for gameplay reasons, some for fiction reasons, some for business reasons, but ultimately they all depend on a handful of key decisions that you make because the solution “feels” right. Those are the ones that make you a designer—an artist, rather than a crafter.

Why would you choose your genre X? Seriously, why would you?

The Old Times

On occasion, people ask me questions about the history of MMOs. Some of these are a touch endearing.

• Q: Why did you write MUD as a text game instead of using graphics?

• A: We didn’t have graphics.

• Q: Do you have any screenshots from MUD’s early days?

  A: We didn’t have screens.

• Q: Were there communities on the Internet before MUD?

  A: There was no Internet.

Technology has moved on so quickly, it’s hard to appreciate exactly what we had to deal with back then.

Did You Know?

Golf was invented in Egypt^{[ 19 ]}, where whatever it was known as hasn’t made it to the present day. The rock tomb of Kheti at Beni Hasan^{[ 20 ]} is adorned with paintings showing scenes of Egyptian life, its east wall focusing on sports. One vignette depicts a man swinging something having the appearance of a golf club at something having the appearance of a golf ball, with the apparent goal of ... well, much the same as whatever it was the guys in the Greek relief were aiming to do.

This painting has also been used to suggest that Hockey was invented in Egypt, although it might have got there from Sumeria where it was called Pukku-Mikku ^{[ 21 ]}. Except the Sumerians used a ring, not a ball^{[ 22 ]}.

Permadeath

Permadeath is a portmanteau word meaning permanent death. It describes the situation in which, when your character dies, it stays dead—obliterated from the database.

It’s hard for modern MMO players to get their heads around the concept of permadeath (or PD as it’s sometimes known). They can’t really grasp the idea that a character is actually >GONE< when it’s killed. Even designers don’t always get it. They’ll come up with ways around it, such as having clones or fighting your way out of the underworld. That’s not permadeath, though. With permadeath, your character Does Not Come Back, Ever.

When the concept has sunk in, today’s players tend to be aghast at the very idea. Their characters can die 10 or 15 times a night when learning a boss fight, and as for PvP, well it doesn’t bear thinking about. How could they play if they kept having to restart from scratch every time their character died?!

What they don’t necessarily realize is that in an MMO with permadeaththat doesn’t happen. We had permadeath in MUD1, and people could and did play for months without dying. They fled from fights they thought they were going to lose. The only times their characters did get killed were when they:

• Left it too long before fleeing

• Attacked some mobile so far out of their league that it one-shotted them

• Were set upon by so many enemies that they fled from one straight into another

Permadeath is a superb mechanic, because it adds meaning. Fights aren’t necessarily more exciting when they involve PD, but people who enter fights knowing their character could suffer PD are of a different order compared to those who will only do it if the worse they can suffer is a wuss slap.

If you can’t lose something, you don’t value it. MMOs are about identity. You only really come to understand how much your character is you when it’s in danger of extinction. Then you know.

Permadeath is never seen in modern AAA MMOs. Although the gameplay payload it delivers is vast, players detest it when it happens to them. That’s the point of it, of course, but it’s also its fatal weakness.

Real life, you may have heard, has permadeath.

MUD’s Setting

Hard though it may seem to believe today, there wasn’t really such a thing as fantasy back when Roy Trubshaw and I wrote MUD.

There was D&D, of course, which explicitly talked about “fantasy campaigns,” drawing its sword-and-sorcery setting from the writings of Tolkien, Howard, Lieber, and (for the magic system) Vance. However, the only other games I’d played that had what we’d now call a fantasy component to them were ones I’d invented myself. As far as the general public was concerned, this was fairytale territory—and fairytales were regarded as being for young children.

As it happened, for MUD’s rewrite as MUD1 I looked at a number of other potential genres. I ultimately went for the one I did (which I thought of as folklore) because it delivered what I wanted in terms of message: I knew that I could speak through it. Although I was only in my late teens, I was fully aware of what this setting could offer because of a board game I’d designed a few years earlier, Wizards & Heroes. I distanced myself from Gygax and Arneson’s interpretation, not because I didn’t like it but because they were saying different things for different reasons.

As it happened, the genre I chose was also the one that would be most enduring. There’s a reason for that: the question, “why fantasy?” has an answer.

Cut from the Same Cloth

On the whole, the argument that 3D graphical worlds are somehow a different species from textual or non-3D graphical worlds is an overstatement. They are certainly different, in the same way that movies with sound are different from movies without sound, but underneath there are so many similarities that they can’t be considered apart. There is such a large intersection between the processes involved in designing and playing both kinds of world that they must be regarded as essentially the same thing.

Great Interfaces of Our Time—2D Profile

IOK, NWN, and KoD displayed their worlds in 2D from an overhead (rendered as near-overhead) perspective. This is not the only way you can display 2D images, though. If you’re going to remove a dimension, it doesn’t have to be height—it could be depth.

The first virtual world to do this was Lucasfilms’ Habitat, by F. Randall Farmer and Chip Morningstar. Here’s what it looked like on a Commodore 64:

To go somewhere, you used your joystick (yes, joystick) to move your cursor over where you wanted to go, then you pressed and held the button and moved the joystick forward (the command for “go”). Then, you released the button. The screen didn’t scroll: if you wanted to move to the region next door, you pointed to the appropriate edge.

By the way, the image above comes from a famous paper^{[ 23 ]} that Chip and Randy wrote documenting their experiences with Habitat. You should read it, if you haven’t already. It’s immensely perspicacious, and has probably had more influence on MMO designers than Habitat itself.

Habitat was innovative but somewhat separate from the virtual world mainstream, which didn’t catch up with the idea of profile-view worlds until much later. Here’s a screenshot from the next one to appear, Steve Nichols’ The Realm Online (launched late 1996). Here’s a screenshot courtesy of Mobygames.com:

In The Realm, each room is a square on a map. You can wander around freely within it, but when you leave you go to the adjacent square (left, right, up, or down—curiously not north, south, east, and west). Again, there’s no scrolling.

Given that The Realm launched just nine months before Ultima Online, which proceeded to wipe the floor with it, you might be forgiven for thinking that this was the end for side-view worlds. Well yes, it pretty well was until MapleStory came along in 2003. Here’s a screenshot, courtesy of Nexon.net:

MapleStory is a Korean MMO with a side-scrolling, platformer kind of movement to it. By April 2006, it had 44 million players worldwide (including 1 in 7 of the population of Taiwan). Okay, so it’s “free,” so those players are “players,” but still!

The Fifth Age—1997 to 2012

The Fourth Age of MMOs was important not because it introduced any new design paradigms to MUDs—it didn’t—but because it showed there was money to be made from them. Almost all virtual world development up until this point had been undertaken by strongly motivated amateurs making up in enthusiasm what they lacked in resources. The regular computer game industry, while aware of MUDs, regarded them as a niche genre; what the Fourth Age showed them was that okay, niche they may be, but HOLY COW, LOOK AT THEIR PROFITS!

MUDs of this era were almost all text-based. It wasn’t that people hadn’t thought of doing (what were at the time called) graphical MUDs, it was that they lacked the money to pay for the artists—and even the computers—necessary to develop them. Text adventures, which had been dominant in the mid-1980s (The Hobbit sold over a million copies), had completely evaporated when graphics came along. Everyone making text MUDs knew they would go the same way once someone was able to get the funding to make a decent graphical one. As it happened, we got three.

The three in question were worked on simultaneously by teams aware of each other’s existence. In order of release, they were: Meridian 59, Ultima Online, and EverQuest.

M59 came out first primarily because it wanted to come out first. It launched in September 1996 in order to claim the title of the world’s first graphical MUD. Nowadays, it is indeed often called the world’s first 3D MMO, although actually it was only 2½D (i.e., it looked 3D but had to fake caves and bridges). Sadly, in its rush to be first, it rather hurt itself: too few people had an Internet connection or a good enough computer to run it, and those who did found a game still very thin on content.

M59 was designed by Mike Sellers, Steve Sellers, and John Hanke. It was inspired partially by Sceptre of Goth and partially by the MUD1 line. Its gameplay was certainly in tune with the times, though, thanks to the influence of a new hire, Damion Schubert. Schubert had been employed on the recommendation of Raph Koster, whom M59 had approached but were too late to sign up: by then, he’d agreed to design a different MMO.

This different MMO was Ultima Online, with which the Fifth, graphical, Age of MMOs truly dawned in September 1997. It needed to attract 20,000 subscribers to break even; in their wildest dreams, the developers speculated that it might get up to 40,000. It garnered 100,000 in its first six months, and over a quarter of a million at its peak. Published by Origin Systems, using Richard Garriott’s much-loved Ultima milieu, its success established MMOs at the forefront of the computer game industry. 100,000 people paying $9.95 a month, none of which was going to retailers? It’s hard today to imagine the impact this revelation had on games publishers.

UO didn’t have true 3D graphics either. It, too, had a 2½D “height map” world, this time rendered in an isometric perspective—a trend followed in Korea by Lineage and still very popular in the Far East to this day. It looked gorgeous, however, and could have held onto its preeminent position had not a combination of circumstances brought it low at just the wrong time. Essentially, it was a victim of its own success: it had more players than it could deal with, both in terms of the demands on its technology and of its customer service. A big change to its PvP system proved deeply unpopular and lots of bad publicity ensued; then just at the worst possible moment, in the Spring of 1999, EverQuest finally launched.

EverQuest, designed by Steve Clover, Brad McQuaid, and Bill Trost, was essentially a rewritten DikuMUD with a graphics engine bolted on. UO had tried (with mixed results) to innovate with new gameplay, but EQ stayed with what had proven most popular in text MUDs—and so struck gold. Within six months it had overtaken UO’s player numbers, and it went on to double them. EQ’s graphics were perhaps more in line with what players of the day expected of a computer game, but it was its combination of visuals with DikuMUD gameplay that made it a winner. Almost all major MMOs launched since EQ have followed the same formula.

The Fifth Age of MMOs was where we were while MMOs strode majestically across the computer game landscape. All that was to change, however, with a simple adjustment to the revenue model.

Ultima Online

Although Meridian 59 was released a full year before Ultima Online, it was the latter rather than the former that kicked off the MMO genre in earnest.

At a time when it wasn’t clear whether (what would soon be called) MMORPGs could be profitable or not, the development of UO was a gamble. The Ultima series of single-player role-playing games was much-loved and respected, so the brand had a good chance of attracting players; however, if it flopped then its developer, Origin Systems, would have taken a big financial hit. Its core team was Richard Garriott (Ultima’s creator), Starr Long (producer), Rick Delashmit (lead programmer), and very soon afterward Raph Koster (lead designer).

UO used a 2D isometric viewpoint, as had its recent predecessors in the Ultima series—particularly Ultima VIII. This meant that some existing code and graphics could be re-used to keep costs down. The world was expansive and looked great, too (at least to 1997 eyes).

As I mentioned just now, the developers had figured they would need perhaps 20,000 players to break even and secretly hoped they might be able to manage 40,000. I myself, when asked at a conference how many players I thought UO would get, was ridiculed for answering “at least 60,000”. We were all way off: it was the first MMO to rack up 100,000 players, and it peaked at around 270,000 in 2003.

Other developers were electrified by UO’s achievement: 100,000 people were paying close to $10 every month to play the game. Whereas before UO, graphical MUDs (as they were called) were all potential, UO realized that potential and opened the gates to riches.

Although its roots were in MUDs, UO brought in innovation after innovation of its own. It had skill-based combat and crafting systems that were components of a vast and complicated economic system. An archetypal sandbox game, UO gave its players enormous freedom to act.

Being a pioneer, though, UO had some unexpected problems. Although it was known from text MUDs that macros could be an issue, the scale of UO meant that macro-ing went on at an industrial rate. Likewise, although griefers in a 2,000-player MUD were controllable, they were not controllable in a world with 50 times that number of players.

The size of the world was also problematic. It was large enough to accommodate (literally, in the case of player housing) the expected number of players, but not the number that actually arrived. Although the economy didn’t care how many players there were, it too suffered: players didn’t behave as expected, leading to resource shortages that seized up part of it entirely. Dupe bugs that allowed players effectively to print their own money didn’t help, either.

Penalties for character death in UO were stiff by modern standards (that is, it had them). Characters had to remain dead for a while or suffer stat loss to come back straight away. Players would probably find their characters’ corpses looted when they got back to them, too. Given that this was a world in which player-versus-player combat was unrestricted, characters tended to die often. A reputation system called notoriety was introduced to warn decent players who were the bad guys, but it gave too many false positives and could itself be used for griefing purposes.

EverQuest came out at a bad time for UO, giving players who were complaining about the effects of unrestricted PvP somewhere else to play instead. In response, Origin created two versions (“facets”) of UO: Felucca was the original; Trammel was the same but with only consensual PvP. Travel between the two was possible, but over 90% of the players decamped to Trammel and stayed there, leaving Felucca practically deserted. This signaled the beginning of UO’s decline. It lost the exhilarating, Wild West atmosphere which was central to its design, and along with it part of its soul.

UO remains to this day a remarkable and influential MMO; there’s still nothing quite like it out there. Its depth and richness, its verisimilitude and its open-worldliness, are matched only perhaps by EVE Online; its inventiveness by nothing. It’s looked back on with great affection by most of those who played it—and indeed those who are still playing it. It’s a true classic; today’s MMOs would not be what they are without it.

D&D and MMOs

A falsehood I often see expressed as if it were a statement of truth is that MMOs are modeled on Dungeons & Dragons.

Well, they’re not.

D&D is but one influence among many on the design of virtual worlds, although of course some MMOs were influenced by it more than others. For example, Ultima Online’s fiction (but not gameplay) owes a debt to D&D, and the DikuMUDs from which most of today’s MMOs descend consciously adopted several D&D tropes—character classes and races in particular. However, they also made wholesale use of some non-D&D ideas, such as a mana system for spells instead of the traditional D&D one-off nuclear option. They weren’t modeled on D&D; rather, they plundered it for ideas.

Even MUD1 did this. It had two specific D&D elements in it: experience points and levels. I looked at a number of possible ways to implement what I wanted to say, and chose the D&D mechanism for character leveling as it was the best fit. The gameplay of MUD1 and the “virtual worldliness” of it were nothing like D&D, and neither were a number of other key features such as character stats and combat. Taking a handful of mechanics doesn’t mean you have a clone.

So D&D did have some influence on MUD1, yes, but nowhere near as much as, say The Lord of the Rings (which also influenced D&D). I used many more ideas from role-playing games of my own invention than I did from D&D, and Roy Trubshaw had never even played it. Few other early virtual world developers had played it, either.

The assertion that MMOs are just computerized D&D does not come from D&D old-timers trying to rewrite history: they know the score. D&D is something special—a fictional form in its own right that isn’t merely a step along the path to computerized role-playing. No, the problem comes from clueless journalists and academics trying to bluff us into believing they’re competent—or from angry players who wish to belittle designers.

D&D is a wondrous game. Treating it as mere step along the way to MMOs insults it.

Why Fantasy?

Around 70% of MMOs are sword and sorcery fantasy. Around 15% are science fiction. The remaining 15% are assorted others (superhero, pirate, horror, modern warfare, whatever). Why is this?

It’s actually better than it was when text MUDs were in their prime, because when we had stock MUD syndrome anyone with access to a Unix box and the Internet could download a ready-made MUD and just run it. The easiest to install and get running were DikuMUD and its offspring, which were almost all fantasy. This meant that perhaps 95% of Third Age text MUDs were therefore fantasy worlds.

It was a different story in the Second Age. Back then, people wrote MUDs mainly for fun, and fewer than half went with fantasy. The mix was rather eclectic, with all manner of idiosyncratic worlds. The fantasy ones tended to attract more players, though, so by the early 1990s, fantasy was dominant. It wasn’t the designers who brought this about, however, it was the players.

Why, though?

Paradigm Consolidation

I wrote this in an e-mail in 2002:

That’s one of my concerns about EverQuest’s being a paradigm: people will come to think that the EQ way of doing things is “right,” even though many of the early changes they made were knee-jerk on behalf of jittery management.

Unfortunately, my concern was justified.

Great Interfaces of Our Time—2D Isometric

Even with pretend perspective, a top-down world view like Neverwinter Nights’ or Kingdom of Drakkar’s is still basically flat. However, if you were to give each square a height, rotate the map so you looked at the squares corner-first, and lower the viewing angle somewhat, you could get something far more persuasive.

That’s what Ultima Online did, and here’s what it looked like in contemporary publicity material:

Behind the scenes, this is still a tessellated 2D squares approach, of course, with a human-size granularity.

Lineage, which was the first behemoth of Korean MMOs, also uses an isometric perspective. It lowers the viewing angle further, which means that in theory if your character is inside a building, you might not be able to see them. However, the client helpfully makes intervening walls transparent so your view is not obscured.

Isometric interfaces to virtual worlds have many attractions, not least that you can make absolutely enormous landscapes quickly and comparatively cheaply. The first one I came across was DragonSpires in 1994, which was developed by just two people (programmer Dr. Cat and artist Manda, both of whom had worked on the single-player Ultima games); they followed it up with the December 1996 launch of Furcadia—a social world in which characters are anthropomorphic animals that’s still going strong almost two decades later.

Isometric worlds can look absolutely beautiful. However, they demand a third-person perspective, which can engender an unfortunate feeling of disconnection with your character when you first start: one more obstacle to surmount when striving for immersion.

Legend

December, 2006: I pick up a copy of Games ^TM, one of the more in-depth UK games magazines, and there’s an article on “The Videogame Legends”. Number five is Richard Garriott.

Now as far as I’m concerned Richard Garriott is indeed a videogame legend. I played Ultimas IV to VI and enjoyed them thoroughly.

Bizarrely, though, the write-up barely mentions the Ultima series. It’s all-out Ultima Online:

Richard Garriott . The so-called “Father of MMOs,” Richard Garriott is responsible for introducing the world to the Massively Multiplayer RPG. Although the genre had technically existed prior to his Ultima Online, he is the man who brought the MMO template to the masses. Ultima Online, and indeed Ultima before it, proved an unbelievable success—more than outselling all its predecessors combined—and established a market that the likes of World of Warcraft, City of Heroes and Guild Wars have since exploited to the max.

When the history of virtual worlds is written, Richard Garriott is going to be regarded as their majesterial inventor, and even those who insist that the first virtual world was Maze War will be unable to do anything about it.

I blame his publicists.

Rank and File

Here’s something you don’t see a lot of nowadays.

Some of the old Star Trek MUDs had a system in which players took on the roles of officers on the bridges of spaceships. There would be one person controlling navigation, another monitoring communications, another in charge of the ship’s systems, another on weapons, another on security duty, and a captain. There’d also be someone down in the engine room to tell you what the warp drive will and willnae take.

When you logged in, you came out of suspended animation and took your place aboard the ship. You could well find yourself as acting captain if the other senior officers were still in bed, and therefore mount away missions or change the ship’s destination if you wanted. Of course, that could mean demotion if the actual captain found out and disapproved.

This mechanism integrates a number of MMO concepts in interestingly different ways. Those officer stations are basically character classes; the ship is a guild fragment; away missions are instance groups; ship-on-ship combat is a PvP raid. There’s great potential to do a lot more, here.

Oh, and if you think this kind of thing is fine for small, hundreds-of-players textual worlds but that it would never scale to a tens-of-thousands-of-players graphical world, check out Puzzle Pirates and change your mind.

D&D Differences

In Dungeons & Dragons, interaction goes through the players; in MMOs, it goes through the characters.

In Dungeons & Dragons, people see the player first and only experience the character through the player’s actions; in MMOs, people see the character first and only experience the player through the character’s actions.

If you like wearing a mask, play MMOs; if you don’t, play D&D.

Six Months of Effort

Typical anti-permadeath reaction: “You mean I could lose my character that I’ve spent six months of effort working on?!”.

Formal response: “No, because you’ll have already lost it several times so you won’t have any characters that you’ve spent six months of effort working on”.

Permadeath is a hard sell.

Mouse Mat

I found this old mouse mat in my almost-as-old laptop case:

It dates from about 2000 and is a screenshot of MUD2 as it looked in the client developed for the Wireplay dial-up service. Sadly, the people who wrote this client only provided an executable and they closed their company down shortly afterward, so no one uses it today. They don’t use Wireplay, either, come to that.

The mouse mat was one of two batches we had made (the first sold out). I kept this one because it wasn’t printed properly—there’s a white line down the left edge where it doesn’t align—so I couldn’t really sell it.

I’d eBay it as important MMO historical ephemera, except I get annoyed when things cost more to post than they’re worth.

Why Fantasy

Science fiction is “about” ideas; westerns are “about” the loner; Greek myth is “about” power; non-game virtual worlds are “about” the community. For some people, those are also fine things for an MMO to be about, but more are attracted to the worlds in which they can become and be heroes than to the ones where they can’t. Fantasy is “about” identity.

It doesn’t matter how much wonder, optimism, iconography, ease-of-implementation, and romance an MMO has, if it’s not about the character then it’s not about the player. Only if it’s about the player can an individual truly feel part of the virtual world.

That’s why fantasy.

When Today Was the Future

Here’s something I wrote in September, 2001:

In 1912, the Motion Picture Patents Company was a consortium that held all the major movie-related patents. It controlled over half the 10,000 Nickelodeons in the USA and was the only organization other than Pathé in France licensed to use Eastman filmstock.

This consortium held all the aces, but none of the founding companies survived the 1920s. They were geared to produce 1- and 2-reel movies only. Independents borrowed the concept of “feature” films from French and Italian film-makers and cleaned up.

Are new games [MMOs] dumping outmoded ideas and forging ahead with something that has the potential to turn the market on its head?

My short answer was no, they’re not.

Damn, sometimes it sucks to be right.

Great Interfaces of Our Time—2½D

Imagine a Chess board. Now imagine that the squares aren’t flat, but are blocks of different heights. Imagine you’re a pawn, standing on one of the lower such blocks, and what you’d see.

That’s a 2½D world.

Okay, so any virtual world based on a tessellation of squares with a height map is, strictly speaking, 2½D. Ultima Online and the other isometric viewpoint worlds are all 2½D. However, they use a third-person viewpoint. The first MMO to give you a first-person view of a 2½D world—a pawn’s eye view rather than a Chess player’s eye view—was 3DO’s Meridian 59, launched in September 1996. Here’s how the Badlands area of it looks with its original rendering engine, courtesy of Brian Green:

Hmm, now you may be thinking “hey, that looks 3D, what’s with this 2½D stuff?”.

Ah, well that’s the thing: it looks 3D, but it’s not. Remember the Chess board? Each square has a height, so you get contours, but what if you wanted, say, a cave at the bottom of a cliff and another directly above it at the top? How would you do that? You’d need for a square to have a height with a gap in it. For similar reasons, you can’t have proper bridges, either.

So, the world is 2D, but the client renders it to make it look 3D: that makes it 2½D.

MUD’s Combat System

This is the story of how MUD’s combat system developed.

Reminder: the program known as MUD1 was actually MUD version 3. MUD version 1 was a proof-of-context written by Roy Trubshaw in October 1978 that he discarded as soon as it ran; it was replaced by MUD version 2 which was playable a few weeks later (before the end of November 1978); MUD version 3 was started in late 1979 when version 2 got too unwieldy, and I took it over from Roy around Easter 1980.

Okay, so obviously version 1 didn’t have anything you could call combat in it. Version 2 sort of did: Roy had planned the too-and-fro exchange of blows that we know and love in today’s MMOs, but he was so busy implementing other components that he didn’t get around to it. I therefore added it myself using MUDDL, the content-creation language that Roy had built into the world. The result was pretty simple, though—basically decided on the flip of a weighted coin. If you lost, you died (permadeath); this didn’t really matter, though, because there was no character advancement in version 2.

When it came to combat in version 3, the way Roy originally saw it was like this: I try to hit you, then you try to hit me, then I try to hit you, and so on, until one of us made contact and killed the other. Yes: one hit and you’d be dead. This is actually a reasonable approximation of how armed combat works in the real world: a succession of minor cuts and bruises you can shrug off until there’s one, incapacitating blow. However, while discussing this design, Roy mused that he did think it was perhaps a little harsh, and that maybe he should give characters a number of lives—3, 5 or 10, say—so they could flee before they got killed forever.

I suggested that we should go with something more fine-grained. Give characters 100 lives, only don’t call them “lives,” call them “stamina”. Give them another property, “dexterity,” to determine their chance to hit and a third property, “strength,” to decide how many stamina points they’d knock off their opponent should they succeed (affected by weapon used). Put in some randomness and some scaling, and the fights would be more exciting: if you were on low stamina then you wouldn’t know for sure if you’d be hit, nor whether that hit would kill you, so you might hang on in the hope your opponent would chicken out first, or you might not risk it.

Now it’s often assumed that MUD’s combat system came out of Dungeons & Dragons, but it didn’t. In D&D back then, characters rolled a number of dice to calculate how many “accumulative hits” they had at their level, e.g., a level 1 “fighting-man” would get one six-sided die plus 1, giving a total number of “hit points” between 2 and 7; a “super hero” would get eight six-sided dice plus 2. In combat, the way you calculated whether you hit your opponent or not was embodied in a table that cross-indexed your level and your opponent’s armor class (from 2 to 9) to give a number; this number was what you needed to roll on a 20-sided die to hit. For example, a level 1-3 fighting-man against an armor class 2 opponent needed to roll 17 or more to hit, but against an armor class 9 opponent would only need a 10 or more. As for how much damage each weapon did, well it was the same for all weapons: 1 point. This changed with the arrival of the first supplement, Greyhawk, which modified the chance to hit by weapon type against each armor class and gave a table for damage inflicted by each weapon. Players were horrified to discover that a sword would do 1-8 damage, especially as the dice rolled to determine your hit points had also been altered (8-sided for what were now called fighters, 6-sided for clerics, and only 4-sided for magic-users and the new “thieves” they’d introduced).

So that’s how D&D combat went. It’s not how MUD combat went. Even today, the core of MMO combat is closer to that of MUD than it is D&D—large numbers of hit points chipped away at by variable amounts in an automated exchange, with individual commands woven in.

Interestingly, it’s the same in most other computer RPGs, too. As these aren’t descended from MUD, all the indications are that it’s just one of those ideas that several people had independently. In D&D it would have been a pain stopping combat to calculate what 16 points off your 73-point total came to, but computers are built to do that stuff.

I sometimes wonder if the concept of permadeath would still be in use if we’d stuck with Roy’s original framing of (what we now call) hit points as “lives”.

Beautiful Architecture

Here’s a screenshot of Meridian 59 extracted from an early magazine review (I believe in Computer Games Strategy Plus, but could be wrong):

Note the caption.

However beautiful your graphics, they’re always going to look poor alongside what comes out later.

Lifespans

I once suggested to MMO veteran Jess Mulligan that characters could have a finite lifetime. This would address many of the issues that would otherwise take permadeath to solve, but would wouldn’t be anywhere near as painful as permadeath.

Jess’s response, mimicking a player, was: “So why would I want to play your game when I know my character is going to be dead after two years no matter what I do?”.

Better in relative terms can still be bad in absolute terms.

Did You Know?

Golf was invented in Rome, where it was known as Paganica.

Golf was invented in England, where it was known as Cambuca.

Golf was invented in Ireland, where it was known as Camanachd.

Golf was invented in Laos, where it was known as Khi.

Hitting a ball into a hole with a stick is an utterly obvious idea. Why wouldn’t golf-like games be invented independently in different countries at different times?

Likewise, having a computer simulate an imaginary world is also an utterly obvious idea. Virtual worlds have been invented independently at least six times: MUD, Sceptre of Goth, Avatar, Island of Kesmai, Habitat, and Monster. We were always going to get them.

Nevertheless, it remains the case that the modern game of Golf is entirely the product of Scotland. Follow the audit trail back from the U.S. Masters and it ends at Scotland. That’s just how it happened.

Similarly, follow the audit trail back from WildStar or Guild Wars 2 and it ends at MUD. That, too, is just how it happened. Writing MUD was not an act of genius.

Okay, so I am a genius, but this isn’t evidence of it.

Evolutionary Arguments

One of the arguments in favor of privileging graphics over text is that humans have evolved to process three-dimensional visual information efficiently.

That’s true, they have; on the whole, men more so than women.

It’s also true that humans have evolved to process linguistic information efficiently; on the whole, women more so than men.

Great Interfaces of Our Time—3D

Yay! At last, I’ve got around to talking about the predominant interface of today’s virtual worlds, 3D.

Aww, you know what they look like anyway, but here’s a screenshot from Sony Interactive Studios’ November 1997 pre-alpha footage of EverQuest, which was the progenitor of fully-3D MMOs:

Okay, so the first thing to point out is that although 3D virtual worlds are 3D internally, they’re not usually 3D on your screen. Your screen is only 2D (unless you’re very rich and don’t mind migraine-category striations across your field of view) (or you’re from the future). A 3D world rendered on a 2D screen is a 2D image. MMO clients can and do present stereoscopic images, which give depth of field on the right hardware; head-mounted displays are still some way away from being a mass-market success, though.

The second thing to point out is that in 3D worlds (and M59-style 2½D worlds, come to that), you can move the camera around fairly freely—as much as the designer decides to allow. This means you can go with one of three basic viewpoints:

• First-person: See the world through your character’s eyes.

• Second-person : See the world, including the back of your character, from a short distance behind your character.

• Third-person: See your character from some distance away, with other characters around having reasonably equal prominence.

“Second-person” here is actually just a short-distance “third-person”. Interestingly, though, the scenery in textual worlds was actually described in formal second-person language (“You are in a vaguely circular room with a high ceiling”).

The third thing to point out is that although you can indeed do caves and bridges in this 3D landscape, the vast majority of it is actually only 2½D. The tools to create the world are mainly about raising and lowering land and populating it with buildings, vegetation, and watercourses. When you do get a cave or a multi-story building, it’s often a prefabricated structure (which is why so many of them have the same layout) or it’s a hand-crafted set piece. This will change as the tools improve, but quite when that will be is another matter.

Of all the ways of representing a virtual world, 3D is the most costly. This is because you need a small army of developers to create it. It’s to be hoped that sometime soon we’ll get a critical mass of free or ultra-cheap 3D models and animations that anyone can use for their virtual worlds, so anyone will use them for their virtual worlds and people can concentrate on improving the way their world is, rather than the way it looks.

It might be a while yet before that happens, though.

Question Me

It’s June 18^th, 2015, the 200^th anniversary of the Battle of Waterloo, and I am looking through the examination papers at Lincoln University in my capacity of External Examiner for their computer game modules. The following question catches my eye:

Richard Bartle famously worked on which pioneering game at Essex University?

• MUD

• Pong

• Donkey Kong

• Tetris

That’s ... scary.

I was so taken aback, it was only hours later that I realized the game names hadn’t been italicized.

EverQuest

Although Meridian 59 (1996) and Ultima Online (1997) were both published before EverQuest (1999), it is nevertheless through EverQuest that the thread of MMO history most strongly runs. There are three reasons for this.

First, it was in the right place at the right time. M59 was in the right place at the wrong time, coming out before the Internet was ready for it.

Second, it only innovated in one area: its graphics. UO also innovated in its gameplay, some of which paid off and some of which didn’t. EQ went with straight, long-proven DikuMUD gameplay that was known to work.

Third, it was a good game that was well put-together. Designed by Steve Clover, Brad McQuaid, and Bill Trost from an idea by John Smedley, it was extensive and enthralling. All four had been players of text MUDs, most notably TorilMUD and Sojourn (which were set in the Forgotten Realms universe of D&D); they learned well from this, constructing the richly-detailed world of Norrath as their setting.

Even by the standards of 1999, EQ’s graphics weren’t great: they were low-resolution, and the windows for chat, character stats, inventory, and the like weren’t transparent. Here’s a contemporary screenshot (with no windows open) courtesy of Mobygames.com:

Nevertheless, in combination with DikuMUD’s gameplay, these graphics were still more than sufficient to attract players to EQ—430,000 at its peak.

Norrath was a tough environment—too tough for many of those who tried it. To advance with any speed, you needed to group with other players: this meant it needed levels of commitment that are today only occasionally present in the hardest-core MMO endgames. This gave EQ a high turnover of players, but while large numbers of newbies were still flocking to the game it wasn’t perceived to be an issue.

EverQuest inspired a generation of MMO designers, and its most famous immediate descendant, World of Warcraft, inspired a generation more. EQ graduates now work throughout the MMO industry. Despite the fact that EQ has its own sequels—EverQuest II (2004) and EverQuest Next Landmark (which entered closed beta in March, 2014), the original EQ is still going strong. Its 20^th expansion, Call of the Forsaken, appeared in October 2013.

As with text MUDs, we were always going to get 3D graphical MUDs: it was just a question of when. That doesn’t mean that the ones we did get aren’t important, though. If someone other than the EQ team had developed a graphical DikuMUD, today’s MMOs would look radically different from what they do today. After all, we know today that the MMOs of tomorrow will be played using brain-computer interfaces, but someone still has to make the archetype of those worlds.

The place that EQ occupies among the pantheon of great MMOs is deserved. Time may have moved on, but class is permanent.

Dingday

In the old days, people remembered when they reached the top level in an MMO and would celebrate it a year on as like a birthday. This wasn’t just a textual world thing, it was the case with Ultima Online and EverQuest as well.

Nowadays, who remembers when they reached the top level in the MMO they play? Who even plays one for long enough to celebrate more than one anniversary anyway?

Why is this?

Remember, we had more choice of what to play back in the text world days, so “next big shiny” shouldn’t be a factor.

Twink

Here’s a term you don’t hear very often these days. A twink is a low-level character geared up with high-level equipment. They were more often a feature of PvP environments than PvE ones, because they looked to be easy prey when they weren’t.

Twinks were invariably secondary characters (or alts as they are known these days—short for “alternate character,” as opposed to main for your primary character). They were common in the past, peaking in popularity in EverQuest. Since then, however, level constraints on items have been introduced and twinks have all but died out. They’ve been replaced by hordes of alts imbued with account-specific perks that enable them to breeze through content to reach the joyous raid-and-reputation grind of the elder game that much more quickly.

Of course, if you play the right MMO then you can just buy a high-level character and never have to touch earlier content again. Or, indeed, never touch it in the first place.

Recovering from Death

If you don’t want your MMO to have permadeath, which these days is almost certainly the case, then you have to accept that whatever fiction you come up with in the context of the MMO to rationalize why you don’t have it is going to feel false unless it somehow explains that player characters are actually immortal. That is possible, of course: characters in Asheron’s Call are brought back to life by artefacts created by Asheron, for example, and characters in the The Secret World recover because they, er, swallowed a bee. However, if you do take this approach then your story can’t treat character death as in any way meaningful—because it’s clearly not.

What you shouldn’t do is waste time trying to think of ever-more-subtle refinements that get closer to permadeath without actually being permadeath, because you’re never going to get close enough unless you actually have permadeath.

The Lion, the Corpse, and the Scratched Box

I wasn’t impressed by the party composition in The Lion, The Witch, and the Wardrobe ^{[ 24 ]}.

Had I been organizing the party, I would have gone for a different make-up. Lucy is a fair enough healer, but Susan makes a dismal rogue: she shoots arrows, but that’s it; she’s more of a ranger than a rogue. She should maybe have chosen a magic-user, she has enough attitude that I’m sure she could hurl the odd fireball if she tried. Peter’s okay as a fighter, perhaps as a paladin if Edmund was also a fighter. However, I’d have maybe made Edmund be a rogue, given his sneakiness, or perhaps have him as a ranger and make Susan dual class rogue/MUser if she bucked her ideas up.

Nah, she has the bow, she should be a ranger/MUser elf; Edmund can be the rogue.

These fantasy authors know nothing about how things work. Sometimes, I think they just make it all up.

MUD and Social Networks

I was once asked by a journalist what debt social networks owe MUD.

You know how astronomers point telescopes to far distant points in the universe and pick up radiation from the big bang that’s still there but is only really noticeable if you specifically look for it? That’s MUD’s relationship to social networks.

There are so many factors involved in the creation of social networks that calling their relationship to MUD a “debt” is wild overstatement. There are plenty of things that MUDs can still teach social networks, but I’d have to be extremely vain to claim that today’s social networks owe MUD or MUDs anything whatsoever.

Why MMO?

Okay, so they’re currently called MMOs, but they weren’t in the past and they won’t be for much more of the future. Here’s a history lesson for you that won’t necessarily make a lot of sense right now, but hey, you could tell this was going to be that kind of book when you dipped inside the covers, right?

So, originally we had MUD. MUD stood for “Multi-User Dungeon,” and it was the name of the great-great-grandfather of today’s virtual worlds. Other pieces of software inspired by it were collectively called MUDs. Some of MUD’s immediate followers used MUG instead of MUD, for “Multi-User Game”. That never really took off, though, not least because it was too general—even something like soccer is a “multi-user game”.

Around 1989/90 there was the Great Schism in virtual world development. We got a split between social worlds and game worlds. The gamers still used MUD to mean all virtual worlds, but the socials used it to mean just the game worlds, referring to the non-game worlds by whatever codebase they used (MOO, MUSH, MUCK, whatever). They used MU* to refer to all virtual worlds, including game worlds—except for those who only used it to mean social virtual worlds.

MUD still held on as the dominant umbrella term until the arrival of Meridian 59. Most virtual worlds before then were textual rather than graphical; there were some with graphics, but these were called graphical MUDs rather than anything special. What M59 brought (or was expected to bring) that made a difference was not the graphics per se but the players attracted by the graphics. Thus, we got “Massively Multiplayer Online Role-Playing Games,” or MMORPGs for short (well, less long).

Now you might be wondering where this mouthful comes from, and because not many people actually know, I’m going to tell you.

Well no, not quite yet, obviously. That would be undramatic.

Public Quests

Public quests are quests that any player in the vicinity can join in when they appear. They were a big selling point of Warhammer Online and a central feature of Rift. They are integrated into the very gameplay of Realm of the Mad God. What’s interesting about them is that they weren’t present in DikuMUD; therefore, their appearance in modern MMOs is a new development.

Well, not so much “development” as “redevelopment”. DikuMUD didn’t have them, but some non-Diku text MUDs did. My own MUD2, for example, has an event called a “mobile bash,” the aim of which is for the players to kill all the mobs in the game world within a single reset (i.e., a fixed time period). The main problem with such cooperative quests lies in fairly rewarding participation. They can be a heck of a lot of fun, though!

Public quests in a modern MMO do have some issues we didn’t in the old days, though, primarily because of the size of their player base. For a public quest to work, there actually has to be a public. There usually is one when a server first opens and a bulge of players works its way through the content. However, if you start six months behind everyone else then you’re going to have fewer people in areas when a public quest starts. The same applies if you’re running an alt. When there’s only you there, this is likely to compromise your ability to complete the quest (unless there’s some ghastly AI balancing system going on, in which case a passer-by or someone who comes to watch can make it worse). In other words, these things are great fun when there are lots of people running around joining in, but frustrating and samey when there are too few people to complete them.

I do agree that putting public quests into a DikuMUD format that has no tradition of them is a good idea, though. That’s if you want formal quests at all, of course.

Peer to Peer

Most MMOs use a client/server architecture, but that’s not the only way you can link computers together. The main alternative is peer-to-peer, which is the LAN game tradition.

As the name suggests, all the computers in a peer-to-peer system are equal—they all have the same responsibilities for running the game. Here’s what a peer-to-peer architecture looks like:

Peer-to-peer is good for multiplayer games, but not for massively-multiplayer games. It’s a distributed architecture, which means work is shared between the computers: you don’t need an expensive server to run a P2P game. Instead, it works by each computer taking it in turn to have control. What you see on your screen reflects the exact state of the game world, because your computer got to decide what that state was at the time it displayed it.

P2P games are easy to set up, and they’re particularly good for people playing on the same nice, fast network.

They have some disadvantages when it comes to massively multiplayer games, though:

• Players need computers that are vaguely similar if no one is to be at a disadvantage.

• If one computer crashes, there can be technical or gameplay issues as a result.

• Games stop when the players stop playing, with no saved state; there’s an inherent lack of persistence.

• Because gameplay-critical decisions are made on computers controlled by the players, the software/data/datastream are open to being hacked. You have to be able to trust your fellow players.

All of these issues can be handled, mainly by massive redundancy—making decisions multiple times on different peers simultaneously talking to multiple other peers. So why don’t we see P2P MMOs?

Well, there’s one more disadvantage to mention:

• You only get to charge people money for P2P software once.

Why Tolkien?

Fantasy as the genre it is today owes its popularity primarily to J. R. R. Tolkien. There were other authors who wrote sword-and-sorcery before Tolkien—Robert E. Howard, Fritz Leiber, C. L. Moore, and Clark Ashton Smith spring to mind—but it is Tolkien who defined it. He created a vast, deep, and self-consistent world—exactly what MMO designers also want to do. Is it any wonder that so many MMOs are fantasy?

Well, it’s not quite that simple. There are other vast, deep, self-consistent worlds out there. The 1,001 Nights are very well-known, and have more magic in them than Tolkien; Robert E. Howard was partly inspired by these stories. Myths from Ancient Greece are also well-known, and are packed full of the kind of things that could make a wonderfully rich and exciting MMO.

Why, then, does Tolkien have such influence? And why is his work still so important for MMOs when they have fireballs and his books don’t?

Different Attitudes

For some people, 80 hours of leveling in an MMO is worth paying $25 to avoid. For others, it’s worth paying $25 to experience.

Sworn Statement

The developers of EverQuest openly admitted to having based it on DikuMUD. That’s fine for gameplay (which can’t be copyrighted, let alone patented), but not so fine for code. Even though DikuMUD was open source, that didn’t mean its code could be used in a commercial product. Members of the DikuMUD community therefore sought assurances from EQ’s developers that they hadn’t used any of DikuMUD’s code in EQ.

Here’s the result^{[ 25 ]}:

As programmer on the server side of EverQuest on we hereby declare that:

EverQuest was not based on DIKU MUD source code, nor is there any DIKU MUD source code in EverQuest. EverQuest is not a derivative of DIKU MUD. The EverQuest source code was built from the ground up. EverQuest has not, in whole or in part, infringed on any copyrighted DIKU MUD material.

I hereby declare that the above is the full truth.

Mr. Brad McQuaid

Mr. Steve Clover

Mr. Roger Uzun

With its assorted errors of grammar, punctuation, and nomenclature, this doesn’t strike me as a great piece of legalistic text; however, the DikuMUD people were happy to accept it and that’s all that matters.

The rumors alleging that EQ contained some of the same spelling mistakes present in DikuMUD continued for a while afterward, though.

Tolkien and MUD

I didn’t base MUD on Tolkien. There isn’t an orc or an elf in there, and the dwarfs are dwarfs, not dwarves. I wanted a world that everyone was bound to, deep-down, that was open-ended, that had mystery and mystique, that was known and unknown, that gave a strong sense of place, that was consistent, that had its own identity. I considered a number of possibilities, but went with what best reflected what I wanted to say: it was fundamentally Northern English, anachronistic and joyous. I used tropes that were straight out of the folk story tradition of my part of the world, because they gave me the symbols with which I wanted to create, to carry meaning. MUD’s “fantasy world” only aligns with Tolkien’s where he used related sources.

This wasn’t a genre I was unfamiliar with, either. As I mentioned in passing earlier, I (and initially two of my school friends) had designed a board game called Wizards and Heroes, which went through several iterations and was very unTolkien. It was pre-D&D, so when D&D came out I D&Dized it to see what would happen; it wasn’t a success, so the next iteration I deD&Dized it (except I kept the priest class, which allowed for more balanced three-player games). The original character of the game, which it retained, was at times evocative of Conan, 1,001 Nights, and Greek myth, and from it I gained a strong idea of the match between the personality of the game world and that of the designer. Thus, when it came to MUD, I sought something that felt right. An author seeks an authorial voice; a virtual world designer seeks an authorial world. MUD was me, not Tolkien.

The Lord of the Rings was a big influence on MUD, it’s true, but because of its world design, not its content: it showed what was possible. It was the sheer scope of it that was inspiring, not the genre nor the fantasy (as opposed to “high fantasy”) aspects. For those, MUD owes far more to the children’s worlds of Enid Blyton and Rupert the Bear. Also, my mother used to make up fairy stories for my brother and me when we were young, which gave both of us an enduring sense of wonder that I tried to infuse into MUD.

Gamer dad, fairytale-writing mum; I was doomed from the start.

Whence the Trinity?

So, combat in MUD1 was pretty much the same as it is in today’s MMOs. There were basically two ways we could have done it: automatically (you start a fight and then it continues by itself until someone dies or flees) or manually (you type a command, it does a hit, then you type another command and it does another hit, and so on). Although Roy Trubshaw and I had discussed the mechanics of individual hits, we didn’t really discuss which of the two combat methods to use—Roy simply implemented the automatic version just before he handed over the program to me. I’m not sure he even considered the hit-by-hit version. I combined the two so that individual commands during combat could be used to mitigate events and thereby add some skill to the proceedings. Fights were a stream of automated events handling the main exchange of blows, but the fun came from doing other things while this was happening so as to increase your chances of winning (casting spells, shouting for help, trying to disarm your opponent, and so on).

This is how combat in MMOs still works, albeit with more spells, more buffs/debuffs, larger fights, but perhaps fewer opportunities to interfere in creative ways with your opponent’s ability to fight (such as trying to weigh them down, or trying to set fire to things in their inventory). That’s at the abstract level, though: obviously, modern MMOs are a lot more sophisticated in other ways and there is some variety among them. One major paradigm-level difference from MUD1 is the fact that today there are character classes that perform different functions—the first MUDs didn’t have character classes.

Now if, back in 1978, you’d told me that there were going to be three main character classes in future MMOs, I would probably have assumed some kind of rock/paper/scissors relationship among them for reasons of balance. Archers beat infantry, cavalry beat archers, infantry beat cavalry—that sort of thing. I don’t believe for a moment I’d have gone with what we have, which is the “trinity” of tank, heals, and DPS (short for “damage per second”—although DD for “damage dealer” is gaining increasing currency). The tank takes all the damage issued by the opponent, the healer undoes this damage, and the DPS gives damage to the opponent. This doesn’t make a great deal of sense in gameplay terms: the healer is redundant (they’re basically just armor for the tank), the premise is unrealistic (“I’ll hit the guy in the metal suit who isn’t hurting me, rather than the ones in the cloth robes who are burning my skin off”), it doesn’t work for player versus player combat (because players don’t go for the guy in the metal suit), and it doesn’t scale (a battle with 1,000 fighters on either side—how many tanks do you need?). Don’t get me wrong, it can be a lot of fun, but it’s a dead end in design terms.

So how did this trinity come about?

Low Levels

MMOs today normally launch with 50 or more levels. Early MUDs had far fewer. The reason for this was because with permanent death you don’t need this many.

As the effects of character “death” gradually diminished over time, we saw a rise in the number of levels that characters could achieve. The points-between-levels began to shorten, too, so that whereas in MUD1 the points you needed to get to level N were double those you needed to get to level N-1, later games had a more logarithmic relationship between their levels. However, overall the number of points you needed to get to the highest levels were far, far more than in the earlier game worlds that had permadeath.

The end result was that in absolute terms, the number of experience points required to go up levels got astronomically larger. In MUD1 you needed 102,400 points to reach the highest level, but eventually we saw textual worlds where the points had to be stored in floating-point number format.

Many concepts are tied to levels. What monsters drop, how much damage weapons do, how much health heals heal, how much armor blocks—they’re all dictated by character levels and they all have to scale accordingly. Basically, when you’re playing at one non-introductory level, it’s pretty much the same as at any other non-introductory level except that all the numbers are slightly bigger.

This is where the “level inflation” we see today ultimately comes from, and why it is that a level 1 character can hit a compliant level 50 character with an axe in perpetuity without causing injury.

Second Life

Given that this is a book about MMOs, and Second Life is not an MMO, it does seem to get mentioned rather more than you might expect.

Well, there’s a reason for that.

MMOs are game worlds, with actual built-in gameplay. Second Life is an example of a social world, with no gameplay. Game worlds and social worlds are subtypes of virtual worlds. Some things about virtual worlds apply to both game worlds and social worlds—their underlying technology is the same, for example. It therefore makes sense at times to talk about virtual worlds, rather than specifically about game worlds or social worlds.

Okay, so the real world occasionally makes decisions about virtual worlds. These could be cultural (“they’re turning our children into flesh-eating monsters!”) or legal (“they must keep a record of every player’s /tell commands because they could be terrorists”) or critical (“they’re not as good as movies”) or commercial (“they’re exceedingly good value for money”)—whatever. Anyway, the thing is, if you’re making decisions about virtual worlds and you know little about them, you need an exemplar. For many people, that exemplar is Second Life.

Now you may ask why anyone would choose Second Life as an exemplar; after all, it’s not a game. Well, there are two main reasons. First, it received a lot of publicity in its heyday and as a result is the only virtual world to have reached the consciousness of many opinion-formers and people in power. Second, as I said, it’s not a game.

Oh, uniformed prejudice.

So what is Second Life, then? It’s a first/third-person 3D world much like any other, but with no game mechanics and no underlying fiction. In their stead, it has the ability for players to create and program their own in-world objects. These objects are constructed from atomic entities called prims (short for “primitives”). Prims come in a variety of shapes (boxes, cylinders, spheres, prisms, rings, and so on) and materials (metal, stone, glass, flesh, wood, plastic, rubber, whatever). There’s an in-world editor to help players cut them up, reshape them, and stick them together. Scripts can be attached to prims to give them functionality.

Now most players of Second Life (or SL for short) don’t actually make anything. It’s time-consuming, you need a certain amount of technical and artistic skill, and once you have made something it’s only the reaction of the other players that determines whether you’ve wasted your time or not.

What most players of Second Life do do is socialize. They hang out together, consume indigenous SL content together, consume content streamed from outside SL together; they organize, they debate, they educate, they protest, they show off—they do all the things that they might want to do in the real world but can’t because it’s real.

So, just like an MMO, then, SL is first and foremost a place. Okay, so it happens to be one that most of the people who tried it found too boring to want to invest time in, but those who do like it tend to like it a lot.

When its star was in the ascendant, SL attracted a great deal of academic research. As a result, many of the paradigms that different academic disciplines have established with regards to virtual worlds in general came out of research into SL. This makes it important for MMOs, especially as a lot of that research was by legal scholars whose work informs lawmakers.

Nowadays, heavy projects dependent on building virtual constructs tend to be done in OpenSimulator (or OpenSim, as it’s usually known), which is an open-source virtual world platform that’s compliant with the SL client. SL itself is declining as technology and fashion overtake it; people new to virtual worlds who like the idea of making things play Minecraft instead. It used to have problems retaining players and making money from them; now it has problems attracting them, too.

Despite all this, SL is still the dominant social world, and is frequently used as a cipher to stand for all of them (much as World of Warcraft is often used as the canonical example of an MMO). It’s already more than established itself a place in the MMO history books—a significant feat for a virtual world that specifically defines itself as not being an MMO.

High = Low x 1.05

High-level content isn’t better than low-level content. It’s the same as low-level content, except the numbers have been multiplied by a factor greater than 1.

Big N Worlds

The first textual worlds were written by players of MUD1. The big three were MUD1, Shades, and Gods, becoming the big four with MirrorWorld, then the big five with Federation II. None of these actually were big compared to what was to come, of course.

The first graphical worlds were written by players of textual worlds. The big three were Ultima Online, EverQuest, and Asheron’s Call, becoming the big four with Anarchy Online, then the big five with Dark Age of Camelot. None of these actually were big compared to what was to come, of course.

Big N worlds show what’s feasible; in the minds of the imaginative, they also show what’s possible. All players want to be designers (“how hard can it be?”), but only a few actually are designers—most just want to play their designs. However, when you get a critical mass of design talent, you have the right conditions for a flowering such as we saw at Adventure ’89.

Dimensions of Inheritance

When constructing game design family trees, you have to decide which primary dimension your tree follows.

For example, consider Ultima Online. This was clearly a direct descendent of Richard Garriott’s Ultima I: The First Age of Darkness, both in terms of the fiction and in terms of the display (UO used basically the same engine and assets as Ultima VIII). However, it was also a direct descendant of MUD1 via LegendMUD, DikuMUD, and AberMUD. So which is it to be? Is Ultima Online properly thought of as a descendent of MUD1 or of Ultima I?

Both views are valid. In one, it’s the Ultima universe with the addition of a new bit of technology; in the other, it’s a virtual world with the addition of a particular piece of IP. Which to go with depends on the dimension of interest to you. If you’re talking MMOs, then UO descends (directly) from MUD1; if you’re talking RPGs, it descends (directly) from U1.

It would be wrong, though, to call a single-player spin-off from UO a descendant of MUD1, or to call an MMO written by UO designers a descendant of U1: the dimensions of inheritance are different.

The Elder Game

When you play an MMO, the “game” you’re playing subtly changes. The core gameplay loop may be much the same at the end as it is at the start (albeit with added complexity), but your goals will have changed. This may or may not be reflected in the design of the MMO, but you yourself will certainly feel it.

The classic point at which there’s a significant change for players is when they reach the game’s level cap. They’ve effectively won the leveling game they were playing (gaining experience points to go up levels), so what do they now do instead?

Well, that’s up to the individual player (quitting is a definite option!), but the MMO’s design will usually incorporate some activities to keep the players in a kind of holding pattern until the next expansion is released and they can start leveling again. These post-leveling, pre-expansion activities form the elder game. It will usually involve some or all of: raiding, instance-running, resource-gathering, PvP or RvR. Naturally, any gains from these will be rendered obsolete the moment the next expansion comes out.

This is the classic form of the elder game, but other elder games are possible. In text MUDs, for example, the elder game could involve being given administrator powers. In EVE Online, the elder game comes up gradually, with individual players themselves making the decision to change their goals to ones that were always there but now seem more possible. Some MMOs in the Far East don’t have an elder game per se: the developer releases new games (rather than expansions), whereupon people leave the now-superseded MMO they were playing to try the new one, regardless of how far they’d got in the old one.

Note that although the terms “elder game” and “end game” are often used interchangeably, there is (strictly speaking) a difference: the former suggests that the player consuming the content will carry on playing; the latter suggests that the player is close to finishing. An elder game can have an end game, but an end game can’t have an elder game.

Should all MMOs have an end game, or only an elder game?

Why MMO

In the mid-1990s, “multiplayer” was something of a selling point for computer games. Developers were slapping the label on anything that could cope with more than one player. This meant that calling a graphical MUD a “multiplayer online RPG” was no big deal—pretty well all online games were multiplayer.

So, there was a UK company called On-Line that developed and operated online games. It started with Federation II, added a dogfight flight simulator kind of game by Kesmai called Air Warrior, plus its own submarine combat game and at one point MUD2. Its managing director, Clement Chambers, said—well, why don’t I let him tell you in his own words?

Back in the day, Total Entertainment Network and Heat.net came out. They were leveraging the success of Doom, and marketing themselves as multi-player games networks providing LAN games over WLAN.

This troubled me because 8-player Doom was not the kind of multiplayer game we were running. TEN and Heat had millions to spend on marketing and were miseducating people to what a multiplayer game was and could be.

I thought we needed a differentiator, as they were 8 players and we were hundreds. The solution in my mind was to tag on Massively to the “multi-player”. MMP was bigger, brighter and better than MP. This of course sounded grand and hit back at the 8-playerness of these new networks.

I put it into a press release and after that, the massively-multiplayer nomenclature got picked up almost immediately.

I know this because Clem used the term “massively multiplayer” on me around two weeks later and, kinda liking it, I asked him where it came from. The above is a recent quote because I didn’t actually remember his original answer and had to e-mail to ask.

Now that sounds cut and dried, but I’ve also heard persuasively that the term originated in a meeting between the developers of Kingdom of Drakkar and those of Legends of Kesmai. Perhaps, as with “newbie,” it’s a word that has been coined independently multiple times.

Anyway, MMORPGs as a term beat another newcomer, Persistent Worlds, to become the most widely-accepted. MUDs became relegated to meaning just textual worlds. Indeed, it’s rare that people today believe they even qualify as being MMORPGs (which is a bit rich, given that even today the biggest textual worlds have more players than some of the graphical ones) (check out http://mudstats.com/Browse ).

You noticed MMORPG was unpronounceable? This eventually became something of an issue. The more tongue-friendly MMOG appeared, but didn’t really catch on for much the same reasons as MUG hadn’t two decades earlier. We then got MMO, which to the untrained eye might appear even more general than MMOG, but as it’s actually an abbreviation of MMORPG (rather than an acronym in its own right), this isn’t the case.

Even MMO isn’t going to last, though. You only have to look it up in Wikipedia, see the legion of definitions of MMO-prefixed terms that some obsessive has defined, and you’ll realize its days are numbered.

For the moment, though, you know what MMO means and I know what it means, and that’s all that we really need from it.

Half a Million Players

From the heyday of EverQuest: “half a million players can’t be wrong!”.

Those wouldn’t include the more-than-half a million who tried it and decided it wasn’t for them?

Whenever people criticize a popular game, the same old “<number> players can’t be wrong!” argument is trotted out in its defense.

Yes, actually: they can be wrong. Thus, in EverQuest’s case, World of Warcraft.

Bandwidth

Ultima Online used to have more external bandwidth than all of the city of New York. Meridian 59 used to have more than all of Silicon Valley.

Limbo

We had a place in MUD1 called Limbo, where wizzes (administrators) put recalcitrant players to cool down. The description read:

• Everything around you is a glowing white, and there are no walls you can focus on. You feel as if you are floating on air. You are.

People in Limbo could talk to other people in Limbo (typically the wiz who put them there), but they couldn’t do anything else. It had no exits, magic didn’t work, and shouts didn’t work. Pretty well all they could do was stew until a wiz released them. They couldn’t even quit, although they could disconnect the phone to get out that way.

As far as I know, no other MMOs (apart from MUD2) implemented the Limbo concept, which I guess suggests that their designers were sufficiently well-behaved never to see the place. Years later, however, Second Life loudly reinvented the concept with its “corn field”. If you hear of something like this in a future MMO, the idea will almost certainly have come from SL, not MUD1—if it hasn’t been reinvented yet again (which it may have been for EverQuest 2’s Drunder “prison server”).

First doesn’t always mean progenitor.

If the Elder Game Is So Good...

Why bother with the leveling game?

We see from the traffic in top-level characters that players are eager to get to the point “where the game really begins”. So why not start there? Why waste time and money implementing a leveling game if people only race through it to reach the elder game as soon as possible? You only need have one level—the “top level”—that applies to everyone. Problem solved!

If the elder game is so good that everyone wants it, why not give it from the beginning?

Can we all go home now?

Grinding

Grinding is when you do the same boring thing over and over again in order to reach the point where you don’t have to do it.

Some players seem to think that grindy content has been put in on purpose to pad out play, with the additional effect that once you’ve been through it you won’t want to quit because then you’d be throwing away all that work. I suppose there could be cynical MMO designers out there who might consider creating grindy content deliberately, but in the past it merely happened by accident. The designers believed that the what-turned-out-to-be grindy content was good content that people would enjoy.

Actually, they could well be right. For long-established MMOs, some of what is now considered grind content used to be “best content” when it was high-end; adding new content beyond it reduced its appeal. This seems to suggest that the best content is merely the content found where all the players are, which is to say the top-most level content. Its actual intrinsic fun value need not necessarily be all that high, but even if it is that won’t save it.

Today’s jewels are tomorrow’s grind. It was ever thus.

The Indie Spirit

Indies shouldn’t try to break into the big N comfort zone, as they can’t compete on polish or content. Someone with $25,000,000 to spend developing an MMO could compete in those areas, but because they can’t afford to slip up they can’t compete on innovation.

The less you have, the less you have to lose. Indies can take risks. Big N gives evolution; independents give revolution.

In history-of-MMOs terms, I’m a dinosaur who isn’t yet extinct. By now I really should be extinct. I want to be extinct! I want to be surprised for once! That’s why I put my faith in indies: it will be an independent designer who takes MMOs in a new, undreamt-of direction.

Oh, to visit such worlds...

Two Unknowns

June 23^rd, 2007. Following a ludium (that’s a special conference-as-a-game thing, etymology fans) at Indiana University, I go into downtown Bloomington and seek out the local games shop. Inside, I find Randy Farmer, who has done the same thing. I buy a game he recommends, Trains Europe, and while I am paying for it Randy picks up a box of World of Warcraft collectable cards next to the checkout desk. “Molten Core Expansion Set,” he reads, then adds, “I’ve never been to Molten Core”. The lad on the till seems pleasantly pleased to find that these gray-haired guys play WoW.

Here’s roughly how the conversation goes:

             LAD

You play World of Warcraft?

             RANDY

Well, I have a level 65, but I've

never been to MC as I'm not in a

raid guild.

             ME

I've pugged it a few times. I have

three 70s, a 60, and a 20-something.

             LAD

I have two 70s. Hey, I think it's

really great that people of all

ages play WoW.

             ME

Well, we have been playing this

kind of game for a while...

Randy glances at me with an "are you going to do it?" look.

             ME

What was World of Warcraft based on?

What game did the developers looked

at and think, "we can do that, only

better"?

             LAD

Er, was it called EverQuest?

             ME

That's right. Do you know what EverQuest

was based on?

             LAD

No, but I think there was some guy at

IU who gave a talk...

             ME

EverQuest was based on DikuMUD,

which was a textual world developed

at the Datalogisk Institut Kobenhavns Universitet in Denmark. DikuMUD

was based on AberMUD, written at

the University of Wales in Aberystwyth.

AberMUD was based on MUD, written at the University of Essex in England. MUD

wasn't based on anything. I co-wrote it.

             LAD

(with disbelief)

You wrote it?

             ME

The first graphical virtual world was

Habitat, written in 1985 by—who

wrote Habitat, Randy?

             RANDY

Randy Farmer and Chip Morningstar.

             ME

We've been writing and playing these

games since before you were born.

I wouldn’t normally do this kind of thing as it’s boastful, but with Randy and me both there together it was just too sweet to pass up.

Making the Impossible Possible

I’m indebted to Mike Rozak for pointing out these promotional screenshots he came across.

So, this detail is from Chronicles of Chronia: Renaissance:

See how those windows cut right across the support beams holding up the walls?

This detail is from SystemHolic Online Rohan:

I love this kind of thing: the pride shown in the texture-mapping of the stones and the modeling of the statues. These are images released to impress the viewer. They largely succeed.

Except, how do those portcullises go up and down?

This is what comes of letting artists build structures in a world that has feeble physics.

Zerging

Zerging is using large numbers of weak characters to overcome a small number of powerful ones. Although the concept has a long history in MMOs, the term itself derives from StarCraft, which features a Zerg Swarm race designed to be used with this tactic.

In real life, zerging isn’t a great approach: you only have to watch the movie Zulu to see why. In human combat, people are alive, wounded, or dead. If they’re wounded or dead, they can’t zerg; if they’re wounded or alive, they can defend themselves. If they have vastly superior technology or a very strong defensive position, you have to break them by siege or go around them (which is what the Zulus eventually did at Rorke’s Drift).

In MMOs, there’s no “wounded” or “dead”. Even dragons can be chipped away at in a death-by-a-thousand-cuts zerg attack until they eventually succumb. Players whose zerging characters are killed in such a fight just step up with their next piece of cannon fodder (which could conceivably be the one that was just killed) and join in again.

Zerging is usually considered as a dubious but begrudgingly accepted tactic in PvP; it’s slightly more respectable in PvE, but not by much. It does have its positive side, in that it can be good for building a sense of community. Participating players may pay for this in self-esteem, though. The group may be strong but the individuals who comprise it, through accepting that they have to be members of the group in order to succeed, are admitting to themselves that they are weak. This can be depressing in a world built to be competitive.

Tibia

Tibia is an MMO released in January 1997, a full eight months before Ultima Online. It’s still running. The reason it doesn’t have much of a profile is that it was developed by a German company, CipSoft, and players from Brazil, Sweden, Poland, and Mexico account for half its player base. Nevertheless, it completed its first decade with around 300,000 active players, which is more than UO managed.

As you might expect from a product of its era, Tibia has 2D graphics. These don’t stack up well against today’s 3D graphical extravaganzas, but they did mean that Tibia (in the form of TibiaME, where the ME stands for “Micro Edition”) was the first MMO to be playable on a mobile phone. Here’s a screenshot from the TibiaME web site:

I mention Tibia because so few developers and historians have heard of it. It’s another example of an underrated virtual world that came chronologically before its more famous peers, yet didn’t have much impact at all on what came after. Sigh.

Victorious Factions

Most realm versus realm, faction versus faction combat never ends. One side can’t completely obliterate the other.

I personally rather like the idea that one side can completely obliterate another. It frequently happens in real life, and in game balance terms it works so long as there’s disintegration pressure on the winning side so it will be riven with internal divisions and eventually split in a civil war. That’s how it happens with EVE Online, which allows players to form their own factions. The other obvious way to do it, espoused by Crowfall, is to have the conquered land itself disappear after a while so you have to conquer new lands from scratch.

Of course, letting one faction succeed isn’t always going to be commercially wise. If Mordor takes over Middle Earth, all that fancy Elf and Hobbit artwork is going to be redundant.

Sixteen Permutations

Speaking of Middle Earth:

• 1 Aragorn

• 2 Arragorn

• 3 Araggorn

• 4 Arraggorn

• 5 Aragorrn

• 6 Arragorrn

• 7 Araggorrn

• 8 Arraggorrn

• 9 Aragornn

• 10 Arragornn

• 11 Araggornn

• 12 Arraggornn

• 13 Aragorrnn

• 14 Arragorrnn

• 15 Araggorrnn

• 16 Arraggorrnn

Hmm, this makes the Arraggornn my daughter saw in Lord of the Rings Online the twelfth least-imaginative Aragorn fan on her server.

The Elroond and Elrroond in her kinship (guild) got along famously.

Alice, Dorothy, and Wendy

Initially, we had Alice worlds. The designers of early MUDs took a sandbox approach—an open-ended world that encouraged players to explore both their new environment and themselves. You were given an ultimate objective (to rack up enough experience points to “win”), but as for how you achieved that, well, it was up to you. The journey was more important than the destination.

The Great Schism gave us Dorothy and Wendy worlds, and left Alice worlds behind.

The designers of Dorothy worlds took a theme park approach—a more closed world that encouraged players to progress through their environment while progressing as individuals. You were given signposts to follow, with all roads eventually leading to a satisfactory conclusion. The destination was more important than the journey.

The designers of Wendy worlds took a content-free approach—a framework world in which players could create new sub-worlds and new sub-selves (and therefore selves). You were given no objective, because people can never fully buy into a fiction that they themselves create. The creation of a destination was itself their journey.

Following the Great Schism, we arrived at the situation that still pertains today: a large number of Dorothy (game) worlds played by a large number of players; a smaller number of Wendy (social) worlds played by a smaller number of players; a much smaller number of Alice (balanced) worlds played by a man, a woman, and their pet dog.

Interestingly, these three world types (plus the empty world) are the very ones predicted to exist by Player Type theory, about which I’ll say more later.

“Interestingly” isn’t really enough, though. Can we somehow turn it into “usefully”?

Silent Movies

Interface aside, textual and graphical worlds are one and the same thing, in the same way that today’s movies are in essence the same thing as silent movies but a different thing from photographs.

Of course time moves on. Fundamentally, though, all virtual worlds have a “virtual worldiness” to them that separates them from other forms of computer software. In the same way that Casino Royale and Fahrenheit 9/11 can both be considered movies, WoW and SL can both be considered virtual worlds. Yes, the film shoots for a feature and (let’s call it) a documentary are different, and there are subsequent design differences, but there’s something about the medium itself which is inviolate and which has been there since before Charlie Chaplin made his name.

So it is with MMOs. Yes, graphical worlds do have different considerations from textual ones, but they’re all part of the same medium. Those design differences don’t go near the heart of what it is for something to be a virtual world; they leave it intact. Modern MMOs may ultimately have got their graphics from Maze or from Spacewar!, but they didn’t get their virtual worldliness from there. They got that mainly from MUD1 and (to a lesser extent) Sceptre of Goth.

Reboot

It’s getting so that the development of MMOs is so expensive and risky that investors won’t touch them.

It’s not just “art house” or back-to-their-roots MMOs that don’t get made, it’s all the rest, too. For every MMO that staggers blinking into the daylight, there are hundreds that die in production—and most of the few that do appear can’t be regarded as having lived up to their promise.

MMOs need a reboot.

Evolution of the Trinity

In text MUD days, physical locations were represented as rooms. These could be of any size, from cupboard to mountain (in the same MMO—scale wasn’t constant, as it didn’t have to be). Rooms were just nodes in a graph. If there were several of you in the same room, then there was no positional relationship between you, you were all just “there,” milling around.

This situation worked well enough for one-on-one fights, but if there were lots of players attacking a single enemy it meant that any one (or indeed all) of them could be hit. This suggested that everyone had to be heavily armored to dull any blows against them, which reduced the ability of players to specialize. Ranged combat or backstabbing-style melée were risky options.

The way such specialization were implemented in Dungeons & Dragons, following the convention of miniature wargames, was based on what would happen in real life: the guys in the best armor would physically block access to the less well-armored characters by using their own bodies. This approach couldn’t work in a world with the room granularity that MUDs had, though, because there was no “space” between the characters and opponents sharing a room.

The solution, which was popularized (and possibly invented) in DikuMUD, was to have taunting commands act to cause the opponent to attack one character in preference to others. Threat-management became a substitute for access-containment. This performed well enough under the circumstances, and added a lot to the gameplay (especially for boss fights); it was a bit of a hack, though. Nevertheless, so it was that the tank was born.

With the tank came the trinity.

When the DikuMUD gameplay was adopted by EverQuest, the trinity came with it. It turned out that players would employ physical blocking to grief one another (e.g., standing in a doorway and not budging), but using the trinity avoided that problem. Almost all of the big MMOs that came after followed EQ’s lead (rather than Ultima Online’s, which was more hard-core), and that’s why we still have the trinity today.

We don’t have to have it, though. There have already been some experiments in allowing physical blocking, most notably in Age of Conan: the grief-your-own-side aspect of it was solved by only performing collision-detection between you and your enemies. However, it wasn’t done in concert with any other trinity-busting activities so it wasn’t as effective as it might have been in plowing a new furrow.

Physical blocking isn’t the only way to replace the trinity with something more flexible and interesting, of course; there are many ways to do it. Hit location would be one way (DPS the legs and it doesn’t move); speed would be another (if it can’t catch you, it can’t hit you); proximity check would be a third (if you’re standing close to someone in armor, they’re deemed to be defending you so they take the hits aimed at you). There’s lots you can do. I personally like physical blocking, but then I also like area of effect spells to damage friend and foe equally, so I’m probably a bit too traditional for most of today’s MMO players.

Anyway, I thought I’d mention this, just so that the generation of players who think their games owe nothing to text MUDs can see evidence of how the latter’s legacy continues (albeit, perhaps, incongruously).

Oh, and just so you don’t think I’m trying to harken back to the halcyon days of yore, I’m not: I want things to advance, not to go backward. If you don’t know where you’ve been, though, advancing is that much harder.

General Ignorance

From The Independent, August 9^th, 2006:

Grandfather? Two years after it launched?

Well, that’s got the family tree sorted, then.

The Sixth Age—2012 to Present

The idea of using microtransactions to fund MMOs is quite an old one. In the West, it was pioneered by textual worlds, in particular Achaea; in the far East, it grew out of gold farming in an “if you can’t beat ’em, join ’em” way. It took off quickly in South Korea and then in China. It didn’t take off in the West, though.

For many years, microtransactions—or free-to-play as the revenue model became known—were rejected in the West. When they were tried, players avoided those MMOs. At the time, this was put down to cultural differences between the West’s focus on individualism and the Far East’s on collectivism.

Although cultural differences were indeed probably important, there were other factors at work. The main one arose from the fact that the longer people play MMOs, the more likely their ideas of what is fun about them are to change. They eventually drift away from the game side of things and toward the social side of things. Thus, whereas in the past they might have regarded paying real money for experience-point boost potions as being cheating, now they no longer necessarily felt that way.

The first mainstream MMO to try switching from a subscription revenue model to free-to-play was Dungeons & Dragons Online. It was decried and derided for doing so, but the results spoke for themselves: revenue more than doubled almost overnight.

Other MMOs followed suit, until in 2012 the dam burst and a whole cascade of virtual worlds dropped subscriptions and started charging real money for certain virtual items. Most of their players now played for free, but some were prepared to pay a little. Crucially, some of the very wealthy players were prepared to pay a mint.

The switch in revenue model brought with it a sea change in design requirements. New MMOs were no longer being created with the aim of giving players lots to do so they would spend lots of (subscription) time doing it. Instead, they were being constructed with the intention of being reasonably playable without payment but with lots of temptations and pressures to spend. The character of MMOs changed as a result, with most old-timers embracing the change but significant numbers—especially those self-identifying as gamers—leaving in dismay at what they regarded as institutionalized unfairness.

The Sixth Age of Virtual Worlds is where we are right now. As to how long we’ll stay in it, well, that remains to be seen. We don’t yet know how the Sixth Age will pan out, nor what will follow it; we can’t even assume it will be welcome. Not all of what’s on the horizon necessarily augurs well for today’s MMO players (or indeed yesterday’s). The Sixth Age, as with any of its forebears, could end in the death of the MMO as a viable concept or in its spectacular rebirth.

Enjoy while you can?

Chain Mail

You have to attach the loose chain vest because that makes this a proper “chain letter”.

Droppable Objects

Objects in MMOs didn’t used to work the way they do now. MMOs had wacky physics models in which dropping an object wouldn’t cause it to evaporate on contact with the ground. Other people could even pick up stuff that you had dropped. Imagine what kind of a crazy universe you’d have to live in if that were possible in real life!

Okay, so this is indeed how the real world (let’s call it Reality) works. Textual worlds worked the same way, because if you want to make players feel that the virtual world is real, it’s easier if said players don’t come across unrealistic behaviors. It’s bad enough that they have to will themselves to believe that magic works; if they also have to will themselves to believe that the world’s mundane physics is magical then that’s unimmersive.

It’s fine for textual worlds to have objects lying around on the ground, but for graphical worlds it’s costly. First of all, you have to store in your run-time database the locations of whatever trash people have discarded, which clogs it up in a way that doesn’t happen if you have a node-based world (as text worlds generally do). Second, someone has to create a model of the object so that you can see it on the ground (at least for 3D worlds).

Both of these are tractable problems. The first one can be solved by having objects despawn after a few minutes. The second one can be solved by having objects that are retexturings of one other or by hiring more artists. At a pinch, you can just have a “there’s something here” symbol that people click on to see (as a 2D icon) what it is.

Being able to drop objects on the ground in textual worlds wasn’t just for reasons of verisimilitude—it had gameplay potential. You could use objects as markers to map mazes, for example, or as traps or bait. If you’re being chased by a wolf, then throwing a dead chicken on the ground might distract it (I know it would distract me!).

Nevertheless, for 3D graphical worlds, the cost of creating models for everything you want players to be able to hold in their inventory is large, so the movement away from being able to drop them was inevitable. Bonus: If you can’t see that the giant moth you just killed was carrying a shield, you’re less likely to think it’s as ridiculous as it actually is.

Importance

MUD is not important because of its commercial success, which was only modest (some $5,000,000 a year into the coffers of CompuServe). MUD is important because of its ideas—because it showed what was possible, because it inspired others, because Roy and I knew what we wanted to see. MUD didn’t take us where we wanted to go (we never expected it would), but it charted the course. I’m disappointed that none of today’s virtual worlds take us where we wanted to go either, but they’re further along the way. One day they’ll arrive, perhaps, although I doubt I’ll live to see it.

People still watch silent movies, but the only reason to make one these days is for artistic or comedic reasons. Nevertheless, the early silent movie makers were pioneers, and their influence is seen everywhere in today’s movies. So it is with MMOs: Roy, I, and others in the 1980s pioneered them, but once civilization arrives then pioneers become unnecessary.

MMOs are to MUDs as talkies are to silent movies, but they’re still in the equivalent of black-and-white. I don’t know what will equate to color, but I do know it’ll come.

Many-Tentacled Beast

There are many MMOs in the Far East. They feel somewhat different from the ones in the West. Surely they don’t descend from MUD1?

Well, they do.

Most modern Oriental MMOs descend from the seminal Lineage, designed by Jake Song. Lineage is the offspring of his earlier MMO, Nexus, Kingdom of the Winds, which itself descends from Baramue Nara, which descends from Jyuragi Gongwon (“Jurassic Park MUD”), which descends from LPMUD and thence from MUD1.

So yes, all those MMOs in South Korea and China do indeed descend from MUD1. MMO history is a many-tentacled beast.

Quiz Quests

One of the innovations introduced by The Secret World is a real-world component to investigative quests. The idea behind this is to support TSW (which has a present-day setting) by seeping some Reality into it and seeping some of it into Reality. An example of the former would be a quest requiring you to know who composed The Four Seasons; an example of the latter would be one in which you need to discover the name of the wife of a fictional TSW character, to be found on his fictional TSW employer’s actual real-world web site.

Of course, the problem with quiz quests is that as soon as one person has the answer, everyone has the answer. TSW’s in-game browser (a nice idea, as they can track your search queries and mine them for problems) will let you Google the question, but (for TSW-bleeding-into-real-world quests in particular) what Google returns first is likely to be a walk-through for the very quest chain you’re following, rather than a reference to whatever information you’re seeking.

Innovation is a matter of context, of course, and needless to say we had some of these out-of-game elements in MUD1. I was against them as a concept, as I didn’t want the real world coming in to the Land (the name of MUD1’s world). Also, puzzles have no replay value: you need to solve one only once then any “fun” from solving it is gone. Furthermore, their answers would soon get around. I might as well just put in a lever saying “pull for treasure” for all the gameplay they added. Still, people did keep asking and asking for puzzles, so eventually I created some just to show how awful the game would be if it were filled to the brim with them. To isolate them from the rest of the Land, I put them in a single, small area: the mausoleum.

The mausoleum had eight doors off it. One was the entrance door and the rest led to tombs. Six had puzzles on them and the remaining (northeast) door was what opened if you got the wrong answer to a puzzle door (the northeast tomb contained a very nasty skeleton by way of disincentive). In MUD2, I added some extra puzzles and randomized which ones appeared so that people couldn’t just go in and press a macro key to open all the puzzle doors at once.

Now the puzzles I came up with to put on the doors were quite varied. Here’s an example:

• ls b sp cl h tcr oc bs ma lg q nhg hp sb wc ea na wa ?

The answer is eb, because those are the initials of tube stations on London’s Central Line going west from Liverpool Street (ls).

Here’s another along similar lines:

• AnEbArPrAyUnUl?

Months of the year without their first letter.

Some puzzles were straight arithmetic, which back in 1980 you had to write a program to solve because the numbers were too big for calculators:

• 2 to the power of 60

• The square root of 6023921858319047472771692203936249

• The 142,812nd prime number

Enjoy.

Some puzzles were in code. This one is easy:

• 1854 151811475 25512121523 7185514 212215 91449715?

It’s a straight letters-for-numbers thing: 18=R, 5=E, 4=D. It’s the colors of the rainbow.

Temporarily diverting though such puzzles may be, the majority of players don’t like them. They find it irritating to be blocked by a smug problem that they know how to solve but which will take them ages. It’s even worse when the answer is obvious once you “get it,” but you don’t get it and there’s no clue anywhere that would hint at it.

Such players’ objection is ultimately the same as mine, in that you need outside knowledge to solve these wretched things. Outside knowledge of a kind is needed in MMOs anyway; there’s a difference, though, between assuming that players know a key will open a locked thing and assuming that they can solve a random second-order differential equation. You can work around most normal MMO problems, but you can’t work around a combination lock with a sequence based on the treble scores on a dartboard cubed or the initials of streets on a Monopoly board converted into numbers in base 16. Logic puzzles and quiz quests are brittle that way.

I kept all the answers to my old mausoleum puzzles. I just wish I’d written down all the reasons why the answers are the answers.

Timelines and Pedigrees

Internet history is often presented as a timeline.

Timelines work well when there are rich, causal connections such that everything affects everything else. In such cases, you can be reasonably sure that anything shown later in the timeline is more or less dependent on everything shown before it.

Timelines are not good when no such causal connections are in place. If you present events chronologically in these cases, you need an audit trail to be able to tie things together in a meaningful manner. A timeline is not a pedigree: merely having been first to the cut doesn’t mean there was automatic influence on what followed.

The concept of the MMO has been invented independently at least six times. However, most of these inventions had little or no impact on the future development of the concept. Therefore, when it comes to understanding MMO history, the important question is not so much which one came first as which of the several ones that came first (as far as their designers were aware) produced offspring that lead to the virtual worlds we have today.

The PLATO games, for example, were certainly prototype MMOs and, depending on your definition of “MMO” and your loyalty to PLATO, some could be regarded as actual MMOs. What influence did they have on the development of today’s virtual worlds, though? Almost none—they’re like Chinese golf.

The second version of MUD had the facility for users to create new objects and commands within the game; that was indeed how new objects and commands were added to it. However, we took that out for the third version (the one that came to be known as MUD1). When Rich Skrenta wrote Monster almost a decade later (without knowing about MUDs), he incorporated the ability to add content from within the game while playing it. This feature was then picked up by James Aspnes when he wrote TinyMUD. If you were to look at a timeline, you would think that TinyMUD got its object-creation system from MUD because MUD had it first; however, the truth is that TinyMUD got it from the much later Monster.

Beware of timelines. Beware of pedigrees too, but timelines are especially good at making it easy to draw false conclusions.

Looking the Same in the Dark

For modern MMOs (ho boy, is that going to sound twee 50 years from now), the virtual world itself is maintained on a server. Players connect to it through a client. The client shows the virtual world to the player from the perspective of his or her character, and allows the player to issues commands for the server to execute on behalf of said character.

Or, if you’re in no mood for stilted, formal language, the client is the player’s window on the player’s character’s world.

Now, many players will look at a client and assume that this is the virtual world. It’s not. It’s just a view of the virtual world. Virtual worlds are maintained on the server, and they’re pretty much the same in terms of how they work irrespective of what their view is like. Most famously, as I mentioned earlier, EverQuest replicated parts of the DikuMUD codebase so closely that its programmers had to sign an affidavit swearing they didn’t include any DikuMUD code in EverQuest. I assume that also means they didn’t use any code from the Sequent codebase, which was the heavily-modified DikuMUD derivative used for Sojourn MUD—the particular virtual world EQ designer Brad McQuaid played.

Thus, although today’s players of 3D extravaganzas may look snootily down on 2½D, isometric, 2D, and textual worlds, actually all these worlds are very similar behind the scenes. 3D clients merely look prettier and give you better motion sickness.

MMO Print Magazines

There are commercially viable newsstand magazines for people who paint wargames miniatures; have blonde hair; collect diecast model cars; fly gliders; own a dairy herd; study the Wild West; are female poker players; want cannabis legalized; keep koi carp; do cross-stitching.

There are millions of MMO players out there. If the cross-stitchers have half a dozen or more magazines to choose from (and they do), why aren’t there half a dozen MMO print magazines vying for our attention? It’s not as if there aren’t free cross-stitch sites online, just as there are free MMO sites.

The first issue of Massive Online Gaming came out in the Fall of 2002. The second issue didn’t come out at all. The first issue of Massive came out in September 2006. Six months and three issues later, it closed down.

Online is best for news, because print takes time. However, a magazine doesn’t have to be a news source. There are magazines for antiques, but it’s not as if there are any new antiques being made. There are plenty of things you can put in a magazine about MMOs. Off the top of my head:

• Comparing the animation styles of WildStar and World of Warcraft.

• The newbie experience in today’s UO.

• What makes RvR tick in Dark Age of Camelot.

• A look at Korean MMOs.

• The free-to-play debate.

• An analysis of tank builds in The Secret World.

• Maps, walkthroughs.

• Interviews with designers.

• How to run a guild.

• Your first raid.

• Children who play.

• Telegraphs: good, bad, or ugly?

• Convention reports.

• Merchandise reviews (books, games, action figures, and so on).

• How to write add-ons.

The reasons that Massive Online Gaming and Massive failed were to do with funding, not the subject matter—and certainly not the quality of writing.

The first issue of Beckett Massive Online Gamer came out in May 2006; the final one in January 2012. Faith can pay off.

A Brief History

MCV (“market for home computing and video games”) is the weekly UK trade magazine for people on the business side of video games (mainly in marketing). You have to be in the industry to receive it, and since I am, I do. Well, I did—it’s all online now,

On Friday, November 26^th, 2010, it had a special, multi-page feature on MMOs. This is how the article began its history of the genre:

MMOs : A BRIEF HISTORY

1975 —The first widely used MUD (multi-user dungeon) adventure game, Colossal Cave Adventure, arrives.

1986 —Kesmai’s Air Warrior for the GEnie online service becomes the first graphical MMOG. The firm’s Gemstone II, the first MMORPG, follows two years later.

1991 —Neverwinter Nights debuts on AOL. It is the first MMORPG to feature graphics. 50 players can log in at once—that rises to a then-impressive 500 within four years.

1996 —Meridian 59 is released. The game is often credited as being the first MMO with 3D graphics.

1996 —Ultima Online debuts, following a series of PC RPGs.

1999 —EverQuest arrives.

Well, that’s my lack of a place in history assured.

I keep telling people in other countries that as far as the UK game industry is concerned, Roy Trubshaw and I never existed. They don’t believe me.

Here’s your evidence, folks.

Trickle-Down

Although droppable-and-pick-uppable-again objects became unfashionable in MMOs, the ability to trade items remained important. Trade, whether through a trade window or through an auction house, is vital if you want your game to have an economy—which you generally do, although they’re not actually essential.

Historically the ability to transfer objects between players was still regarded as being A Good Thing even in MMOs without a complex economy. It meant that low-level characters could have access to better items when high-level characters upgraded. You would sell your +4 sword once you got your +5 sword, so a lower-level character could use it. This was called trickle-down, and was seen as a way of redistributing wealth.

It wasn’t without problems, though.

30 Up

October 20th, 2008: MUD’s official 30^th anniversary.

It was reached without causing a ripple of interest. There were no articles in newspapers, no radio interviews, no podcasts, no blogs: the only people who noticed were the players of MUD2 (who held a MUDmeet).

Why was this so?

Well, the mainstream media have no interest in anything but the new. That’s why they call them newspapers. Computer games are new, MMOs are newer, so a reminder that they were invented 30 years earlier didn’t sit well with the narrative.

As for the game industry, well, some old-timers know the history of MMOs and whence they came, but most of today’s developers haven’t a clue, nor do they feel the need to get one. They can always reinvent anything old and they see no reason to know why things are the way they are, just that they are the way they are. The ones coming out of college know all about Pong and Atari, but MMOs are just niche insofar as history goes. Also, they were unfashionably invented in the UK, not the United States or Japan, so don’t slot neatly into a timeline (and even when they are included, there’ll be 50 entries ahead of MUD).

MMO players themselves were unaware that it was 30 years since their hobby began. I occasionally ask those I meet in-world if they know its development history. Few are even able to cite EverQuest as an influence, although occasionally someone will have a stab at “Ultimate Online”. With even the graphical worlds of the early 2000s forgotten, what hope do their textual precursors have? Older players have sometimes played MUDs, and recognize the connection between those worlds and today’s MMOs; few know which one started it all off, though.

It’s easy for me, who every once in a while is lauded as the father/grandfather/great-grandfather of MMOs (which I’m not—you want Roy Trubshaw, not me), to get the impression that MMOs are a major cultural influence of the 21^st Century and that I’m a minor celebrity. In truth, though, neither they nor I are anything of the sort. MMOs will, I believe, finally reach their potential—but they haven’t yet. As for me, well as I’ve said before, the idea of having a computer simulate an imaginary world is obvious—we were always going to get them. The fact that almost all today’s MMOs descend directly from MUD1 rather than Sceptre of Goth, Avatar, Island of Kesmai, Habitat, or Monster is mostly a right place, right time accident.

So standing back and looking at it, the answer as to why there was not a lot of fuss over MUD’s 30^th anniversary is that in the great scheme of things, it wasn’t actually important. The mainstream wasn’t interested because MMOs haven’t had much impact; developers weren’t interested because the paradigm is obvious; players weren’t interested because knowing doesn’t add anything to their play experience; academics might have been interested in the historical facts, but anniversaries don’t figure in their analyses.

The only people who could be expected to be interested were those who played MUD2, because MUD is their world. As their celebratory MUDmeet showed, they were indeed interested. Almost everyone else may be playing worlds that descend directly from MUD, but they’re not playing MUD; their worlds have gone in different directions.

Which, in the end, is as it should be.

PLATO Revisionism

Every once in a while, I’ll be contacted by someone I’ve never heard of before who is writing a book or article on the history of computer games. I’m happy to cooperate, because so much is misreported about their history that it’s possible for an entirely false picture to be built up if people don’t do their research. Providing that the person asking doesn’t want so much information that I may as well be writing the piece myself, I’m glad to help.

Occasionally, however, the people who contact me aren’t trying to obtain or verify information, they’re aiming to fit an agenda. For some reason, authors writing about the early and highly influential PLATO system at the University of Illinois are the most revisionist in this regard. The vast majority of game historians obviously aren’t revisionist, but of the ones who are, most in my experience have been writing about PLATO.

I believe the reason for this is that PLATO was the root of so many advances in computers (plasma displays, touchscreens, message boards, flight simulators, Freecell) that it’s easy to assume it to be the root of everything. It certainly had more influence on computer games than one might expect, because of its use of graphics at a time when everyone else was using text. However, its influence on MMOs is minimal.

I do know some industry old-timers who cut their teeth on PLATO and who have influenced MMO development for the better; in particular, Gordon Walton, David Shapiro (Dr. Cat), and Andy Zaffron have made big contributions. Overall, however, the PLATO system had only a drop-in-the-ocean effect. Today’s MMOs are descendents of the textual worlds invented in the late 1970s and the 1980s—primarily MUD1 and to some degree Sceptre of Goth. PLATO did have its own (what we’d now call an) MMO, Avatar, but despite its huge popularity on PLATO it never jumped species to propagate outside of PLATO; any evolutionary path that derived from it was ultimately moribund.

I do get the impression that some ex-PLATO people (none of those mentioned above, I hasten to add!) can’t quite get over the fact that there are some aspects of computer game history that don’t have PLATO as their starting point. They like PLATO and they want it credited for its many numerous achievements, which is fair enough; it’s not fair, though, when they start claiming credit for things it didn’t pioneer (or, if it did pioneer them, they fell by the wayside). MMOs are one such thing.

There was a game on PLATO called Oubliette, which predates MUD1 by about a year. A while back, I managed to play this on a PLATO simulator. In terms of its “virtual worldliness,” I wasn’t impressed: its inter-player communication was extremely limited, its inter-character interaction was nominal, and its world persistence was zero. It was close, but not close enough for me (you might be more forgiving, of course). In my view, it’s a multi-player game, but not quite a virtual world. Avatar, on the other hand, is definitely a virtual world by any standards. Both, however, fall into the “golf was invented in China” category when it comes to MMO history.

If you think I’m being unreasonable, here’s what the opening of the “history” section for the definition of MMORPG in Wikipedia used to say:

The beginning of the MMORPG genre can be traced back to non-graphical online Multi-User Dungeon (MUD) games such as those developed in the late-1970s for the PLATO system. Earlier games such as pedit5, dnd, Dungeon, orthanc, baradur, bnd, and sorcery were multi-user games, but the players could not interact with one another. Subsequent games on PLATO, including oubliette, avathar (later renamed avatar), emprise, and moria allowed players to interact, including helping each other in battle and trading equipment.

You can probably trace Avatar back to pedit5, but Avatar’s and World of Warcraft’s paths don’t cross. Given that most players can’t even trace World of Warcraft’s lineage even as far back as EverQuest, though, does it matter if it’s traced back to pedit5 instead of MUD1?

Well, if history matters, it matters. If history doesn’t matter, it doesn’t. If it didn’t matter, though, why would some people wish to change it?

Trickle-Down Problems

I mentioned that trickle-down as a concept had problems, which is why we don’t see it advocated much today. Four in particular are worth examining.

First, it was possible to give extremely powerful objects to low-level characters (the twinks I described earlier). They could cut down monsters that were supposed to be challenging as if they were made of butter. What’s more, they could do the same to other characters of their own level in PvP. The solution adopted for this was to put limits on what objects players could use, so a sword intended for level 50 content couldn’t be wielded until you were level 40, say.

I personally don’t like this approach: if you can wield a sword, you can wield a sword—the level of the sword shouldn’t make much difference. What does make a difference is your degree of skill. I’d have liked instead for a level 10 person wielding a level 50 sword to get only a minor advantage (because it’s sharper, say, but they’re not skilled enough to exploit this properly). However, by this point in MMO history the ill-conceived shift of stats from character to equipment was well established, so my preferred solution wasn’t an option. If you’re only as good as your gear, then to stop people from becoming too good you have to limit their access to too-good gear.

The second problem with trickle-down arose because characters accumulate at the high end. This meant there were more players selling high-level things than there were buyers for them. Coupled with the inflation that typically accompanies levels (1GP at level 20 could seem a fortune, but at level 60 could seem a pittance), this meant the higher-level objects were practically given away. Those high-level players who felt that their uber-elite armor ought to be worth more than next-to-nothing were annoyed by this. Anyone could have over-powered gear if they sought it out.

To fix this properly, gear should have been damaged more in combat (it’s usually combat gear we’re talking about here) and have a half life before it disintegrates beyond repair. You can imagine how popular that would have been among the players, though. The introduction of bind-on-equip and bind-on-use mitigated this by allowing players to sell stuff they found but didn’t need, while preventing them from selling things they had used but no longer needed. It’s a hack, though.

The third problem with trickle-down was that sometimes it was trickle-up. In worlds with meaningful penalties for getting killed, why would you send your main character into a dangerous zone when you could send in an alt whom you didn’t mind getting killed? Having died on your alt a few times obtaining something stealthily, you could then pass what you got on to your main. This doesn’t happen so much now, though, as MMOs have relaxed death penalties to slap-on-the-wrist level.

The fourth and final problem with trickle-down was trickle-sideways. Some objects in MMOs will be very powerful and rare. You can only get a shot at obtaining them every few days; success is not guaranteed, and even if after four hours you down the boss that may drop what you want, there’s a good chance it won’t. Ah, but what if someone else already has the item you want? They could give it to you! Hmm, but why would they do that if it’s so powerful? Well, they might exchange it for real money.

What happened, then, was that people started to farm rare items in order to sell them to other players. This increased the supply of the rare items (adding some imbalance, although not show-stopping amounts) and also introduced competition to fight the bosses that dropped those items. There could be scores of people camping a location, waiting for a boss to appear so that they could attack it. Most of those people would be farmers, who were in effect imposing a toll on the regular players: if you want this rare object, that you could get far more easily if we weren’t here, you’ll have to pay us for it. Shifting the bosses to instanced content removed this camping aspect of the problem, but it meant that farmers no longer had to compete with each other—each group could have its own, private boss to kill. This greatly increased the supply of rare objects, which then sometimes did create imbalance by a show-stopping amount.

Sceptre of Goth Alumni

I always say that “almost all” today’s MMOs are directly descended from MUD, because if I said “all” it would be untrue. A good few are direct descendents of Sceptre of Goth, and some former Sceptre players are among the most influential in the MMO industry.

Perhaps the most famous of these is Mark Jacobs. As a young, ardent Sceptre player, he applied for a job at InterPlay but was turned down. When InterPlay then closed his local Sceptre franchise he snapped and set up his own company, Adventures Unlimited Software Inc., to create a rival game that became Aradath. Following Aradath, Mark designed Dragon’s Gate, one of the big-hitters of the Fourth Age of MMOs. He went on to design Dark Age of Camelot and Warhammer: Age of Reckoning among other titles. He’s a towering figure among MMO designers, hugely respected for his passionate support of the ordinary player.

Matt Firor and Rob Denton had tried to buy a Sceptre franchise, but InterPlay’s collapse stymied them. They set up a company, Interesting Systems Inc., which later merged with Mark Jacobs’ AUSI to form Mythic Entertainment. Electronic Arts acquired Mythic in 2006 and merged it with Bioware in 2009.

• Firor left in 2006 to become president of ZeniMax Online, developers of The Elder Scrolls Online.

• Jacobs left in 2009 and set up a new company, City State Entertainment, to develop Camelot Unchained.

• Denton stayed on as Mythic’s General Manager and headed up Broadsword Online Games, a studio created to develop, support, and operate DAoC and UO for EA.

Game designer and Sceptre player David Whatley set up Simutronics in the aftermath of the demise of InterPlay, along with former InterPlay vice-president Tom Zelinski. Simutronics enjoyed great success with its GemStone series of textual worlds, which still operate today. Simutronics is not just about text, though: its Hero Engine is widely admired and was chosen by Bioware for developing Star Wars: the Old Republic.

Scott Hartsman played on the Milwaukee Sceptre franchise. He was one of several ex-Sceptre players to work on GemStone for Simutronics. Scott later moved on to Sony Online Entertainment, becoming EverQuest’s technical director, then senior producer and creative director for EverQuest 2. He finally hit pay dirt as CEO of Trion Worlds, the developer of Rift; I have to say, it couldn’t have happened to a nicer guy.

The list goes on. Andrew and Chris Kirmse formed Archetype Interactive to develop Meridian 59. Both were former Sceptre players. Wolfpack Studios, which developed Shadowbane, was set up by several Sceptre veterans, including Todd Coleman (who went on to make Crowfall). Brett Vickers, the lead programmer for Guild Wars 2, cut his teeth on Sceptre.

I’m telling you all this for three reasons.

First, it shows that the history of MMOs does not just begin with MUD. Had InterPlay not gone under, we could now be calling old textual worlds Sceptres. I greatly dislike it when people try to take credit that is not their due, but I dislike it more when people are not given credit that is their due; credit is due Sceptre’s pioneers, so I’m making sure you know that.

Second, if you know the names of designers, you know that MMOs are designed. You understand that each designer is different, with a different background and different things to say. You appreciate that MMO design is an art form.

Third, the reason these people are the important figures they are today is that they were so driven by a desire to create their own worlds that they made it happen. Sure, you may rather like the idea of building your own MMO, but do you want to do it so badly that you’re prepared to throw away a promising career in law and learn programming from scratch to realize your ambition? Mark Jacobs did.

That almost all today’s MMOs are directly descended from MUD is in part down to luck. That doesn’t mean that if luck doesn’t favor you, you can’t make it to the top anyway, though.

Avast, Tharr, ye Avatars!

In 2001, I did some consultancy sessions for a pirates MMO (it was canned, so I’m allowed to talk about it). The idea was, the player took on the role of a ship’s captain, controlling the ship while at sea but the character while ashore. This meant the player was represented by an avatar while wandering around the pirate towns, but as a ship while on the high seas.

An interesting thing about this was that although you could sink ships, the captain (that is, you) always managed to escape. Thus, although your character could never suffer permanent death, many of the beneficial effects of PD were nevertheless present if your ship-you was at the bottom of the ocean. You had to do the start-from-scratch thing, but didn’t have that “oh god oh god oh god I’m DEAD” moment that can make PD so inordinately distressing.

Players of the space opera EVE Online will recognize this mechanism, as it’s employed there. You’ll generally survive as an individual (in an escape pod, or, if that gets shot up too, as a clone), but the material costs of defeat are somewhat more biting than in the likes of WoW.

The usual way to avoid implementing PD is to state that a character is dead when it isn’t. “He killed me!,” yeah, right. As EVE shows, though, you can also avoid it by doing the complete opposite: state that a character isn’t dead when it is.

Ever Forward

What I want is for MMOs to become what they have the potential to become, not to rebecome what they once were. When I started on MUD, back in the late 1970s, I had to look forward because there was no backward to look; I still look forward, and I always will look forward.

When people today take a backward step, I point out what they’ve just lost. This may appear to be nostalgia for past glories, but I don’t see it that way. For me, it’s trying to drag MMOs kicking and screaming into the future, not miring them once more in the past.

BOA Constriction

In response to the problems that item-farmers introduced into MMOs, highly-desirable objects were made bind-on-pickup.

The idea was that BOP items couldn’t be traded. If someone had one, the only way they could have got it was to have killed the boss that dropped it. This restored the supply to its intended levels and removed at a stroke the quite legitimate complaints coming from some players that other players were buying success.

Bind-on-account was an innovation that allowed players to transfer goods to their own twinks, but not to those of other players. It breaks the fiction of a world even more than BOP does, as an account is a concept from a reality external to that of the virtual world itself, and is therefore not something meaningful within it. You do occasionally get some scant cover, such as the legacy system in Star Wars: the Old Republic, but it’s never natural.

My own view is that bind-on-pickup/equip/use are necessary evils in the current MMO paradigm, but that the current MMO paradigm is in dire need of an overhaul. Putting the equivalent of biometric passwords on swords makes no sense. Putting it on armor makes more sense because properly, any given piece of armor should only fit people who are roughly the same size. The way it’s implemented at the moment, though, armor either fits you so well that you can’t even give it to someone else (BOP) or it will fit them even if they’re a female gnome and you’re a male orc (because BOE makes it conform to the shape of the first being to don it).

If you want to visit another reality, it has to feel like a reality. It has to make sense. BOP, BOE, BOU, and especially BOA interfere with this, which is why ultimately I’d like to see them all go.

It’s going to take a paradigm shift, though.

Prompt Response

MUD1 was textual in nature, which meant you typed your commands at a keyboard. Trust me, graphics don’t work at all well on a 110-baud teletype printing uppercase letters onto fanfold paper.

For those of you who have never used a command line, the way it works is that the system prints a prompt to let you know it’s ready to accept input. You can usually type ahead (very handy during periods of heavy use when you’re watching output appear at six characters per second), but if you do then you can’t see what you’re typing—or you can, but it’s interspersed with the output. You might therefore want to wait until there’s no more output pending before you start your next command. The appearance of the prompt tells you it’s okay to go ahead: the system is waiting for a command.

Different programs used different prompts. The TOPS-10 operating system upon which MUD1 was developed used a dot prompt, for example; the BASIC programming language interpreter used the word READY.

When you’re writing for a teletype, you really don’t want a long prompt, so MUD1 used a single character, a star: *. It had a different prompt when you were in wizard mode, though, to remind you that you were in wizard mode. Players might get upset if you forgot and attacked their character, as it would lead to its instant and permanent death.

The wizard mode prompt was the same as the regular prompt, except it was preceded by four minus signs. This meant it looked like a magic wand (on a teletype at least).

This is why, when I wrote this book, I used ----* to separate the articles. You can thank the copy editor for being spared the result.

Elder Game Emptiness

At the moment, the MMO paradigm is to have two games: the leveling game and the elder game.

As I’ve already explained, the leveling game is where you rack up experience points for doing stuff, causing you to go up levels and gain new abilities. Eventually, you reach the level cap. The “game” ought to stop there, but it doesn’t because developers are scared that they’ll lose players if they let them feel they’ve “finished”. Thus, we get the elder game.

The elder game itself can involve many activities, but primarily comes in two flavors: the raiding game and PvP/RvR.

In the raiding game, you and your guildies spend a week in an instance dancing with buggy bosses for a shot at the last piece of top-of-the-range gear you need to complete a set that will be superseded in the next major patch.

In PvP/RvR, you spend your time fighting other players for tokens that you can cash in for gear that will enable you at last to beat the people who, if they hadn’t beaten you, you would have got tokens for beating. There’s a PvE variant where you collect tokens to cash in for tokens that help you collect tokens better.

Neither of these elder game alternatives has much at all to do with the leveling game. Both only put the player on hold ready for the next leveling game, which they will then consume like locusts until they’re back at the elder game and are bored again.

I asked earlier why, if the elder game is so good, do we need the leveling game? Well, it’s because actually the elder game isn’t good. Raid content is fixed and too samey. Realm versus Realm is never resolved and is therefore pointless. PvP has a positive feedback loop, so the worse you are, the harder it is to catch up.

What we need is an elder game that is meaningful, in which players can decide for themselves what’s fun instead of passively expecting designers to identify it for them.

Something that could help sort that out would be interesting. More than that, it would be useful.

Indie Development Costs

The development of early textual worlds was hindered because mainframes cost more than ocean liners. Only when PCs came out could home users think of writing MMOs. Costs were still high, but by 1989 they were tantalizingly within reach. Here’s a clip from an Adventure ’89 advertising flier:

The development of early graphical worlds was also hindered by cost: hardware, bandwidth, and artwork were the main worries. They remain expensive, to run as well as to develop, but their costs are nevertheless tantalizingly within reach. You’ll make it all back once you launch, right?

Back in 1989, even small-scale textual worlds made money. Here’s why (from a MUD2 flier I handed out at the same event):

MUD’s pricing structure is based around the “credit”. Credits are units of time on the game (each one represents 12 minutes). As you play, your credit total is automatically reduced, five per hour. Credits are bought in advance. The more you buy, the less they cost. The cheapest rate works out at 50p an hour and the most expensive is £1 an hour.

£1 of 1989 money is over £2 of today’s money. Imagine if nowadays you could charge players £2 an hour—or even $2 an hour—to play an MMO! Back then, people figured that these game worlds would make money enough to fill several barns. Indeed, some actually did make life-changing amounts of money, but most just about broke even. Developers innocently believed that their game was their product, whereas actually it’s the company that is the product; the game is merely what the company sells in order for it (the company) to be worth something.

Assuming you have a brilliant idea for an MMO that everyone will want to play, there are two ways to make eye-boggling amounts of money from it:

• 1. Spend eye-boggling amounts of money on development.

• 2. Be lucky. Indies tend to aim for this one.

For indies, it’s easy to persuade yourself that just because you build an MMO, players will come, and that originality means success. Unfortunately, the stark truth is that if you aim for half a million players you will not get them. That’s not. You can regard 100,000 as a cause for unfettered celebration and 40,000 as a major achievement; after all, you probably “only” need 20,000 to 30,000 to break even.

However, you don’t actually care, do you?

Indies don’t develop MMOs to make money. They develop them because they have to—it’s their medium of expression. It’s who they are. This is why if you don’t aim for half a million users, you can still wind up with many times more than that number, as happened with Minecraft and DayZ.

Indie development costs, but it also rewards.

Type Systems

Type systems in programming languages are supposed to empower programmers with the ability to portray abstractions, protect them from making mistakes, help with optimization, and assist in documentation.

Speaking as an old-time hacker, this is depressing stuff. For me, type systems limit programmers their ability to portray abstractions, prevent them from making mistakes they were never going to make, help optimize code that might have run faster if the programmer could have had stronger control in the first place, and assist in documentation no more than might a decent choice of variable names.

You can say anything you like in an untyped language. In a typed language, you can only say what it allows you to say. Never mind modern languages such as C# and C++; to me, even their common ancestor, C, is over-typed. I grew up with BCPL, which is typeless (or, if you like, it has one type, the word; or, if you really like, it has an infinite number of types, just no coercion).

What’s needed is data structuring, not data typing.

What You Need to Know to Write an MMO (1978)

Memory is made of soft-iron toroidal cores suspended across a grid of wires, with a diagonal sense/inhibit wire to read/write them.

Circuits are made of wires and gates. An and gate is like two switches in series; an or gate is like two switches in parallel. By combining and gates and (rather trickier) not gates, you can make nand gates, from which you can construct any logic circuit. One such circuit is a flip-flop, which has two stable states and so works as another (more power-hungry) form of memory. You can connect flip-flops in a chain to make a register. By passing the inputs to these flip-flops through a set of gates, you can effect operations—shifting a register’s bit settings all to the left, for example (which is the same as multiplying the number the register represents by 2). Another arrangement of gates makes a circuit called a half-adder; combining multiple half-adders will allow you to add the contents of two registers together.

In order to decide which operation to perform on an accumulator register, you have a special instruction register. By using gates to detect combinations of bits in the instruction register, the operation it identifies can be determined. Instructions are stored in memory, and the address of the next instruction to fetch is kept in a register called a program counter. Having fetched an instruction, it is executed, and then the program counter is incremented so the next operation can be fetched. Operations on the program counter, such as addition, implement concepts such as jumps.

A set of panel switches load initial values into registers. This is tedious, so just enough are loaded to enable further instructions to be read from paper tape. These allow many more instructions to be loaded into memory from a magnetic tape. The program loaded from the magnetic tape is the operating system, which will administer the computer (for example, manage the devices that application programs run).

User programs are kept on disc packs, cards, paper tape, and magnetic tape. To write a program, you use a programming language. The first programs are entered using the panel switches; they ultimately create a program called an assembler, which converts human-readable assembly language into the binary that the computer actually executes. Assembly language is used to write the makings of a compiler for a high-level language. Compilers take programs written in high-level languages and drop either assembler or direct binary. Just enough of the compiler is written in assembler that the rest of it can be written in the compiler’s language itself, a piece at a time across multiple iterations.

High-level languages can be used to write general-purpose programs. Almost every program is some combination of the activities performed by a compiler, a database, and/or an operating system. You need to be fluent in a high-level language before you can program much in it, but they’re all basically the same when it comes down to it. Choice of language is a balance between speed of execution, speed of writing, and speed of maintenance.

To create a virtual world, you need to design your own data definition language to specify the virtual world. You write your own compiler to convert it into assembly language plus some intermediate code that you can interpret (that is, execute in software rather than in hardware). You store details about the players and the world itself in a database, which you also write yourself. When the virtual world runs, it acts like an operating system—continually processing until it is stopped, crashes, or exits gracefully.

If, when you create your virtual world, you don’t know the whole story from memory cores (or their transistor equivalent) right the way up to write-your-own-compiler level, you won’t be able to tweak every bit, stretch every instruction, pack every data structure or take every shortcut; less of your world will fit in the (in MUD1’s case) 70K of memory you get in evenings and weekends on the local timesharing system. You need to know all this if you’re going to do it.

Oh, you also need that little bit at the end where you design the game world itself.

Company Organization

Regular non-game software engineering companies are usually organized along functional lines:

• Company leadership

• Sales and marketing

• Finance and accounting

• Software development, support, and quality assurance

• Operations and IT

• Human resources

There are variations of course; for example, there may be both a sales department and a marketing department, but the above structure is fairly typical.

Software development creates the products; the other divisions support the sale of these products and the running of the company itself. There won’t usually be a group for product specification, though—the specification task falls to whoever sources the software. For large companies (such as banks) that have in-house programming sections, there will probably be a business development unit that does the specification.

Regular game software engineering companies (which the industry calls developers, or more formally, development studios) have a modified version of this standard organization:

• Add an art and animation section

• Add audio (music and sound effects)

• Expand quality assurance and separate it from software development

• Include a small design group

There is variation here, too; audio may be outsourced, for example.

Regular MMO software engineering companies take the modified game developer model and refine it further:

• Expand the operations section

• Expand the design group

• Greatly expand the support section and separate it from software development

• Reabsorb quality assurance into this new support section

Variation here includes combining operations and support, and some differences in the make-up of the individual sections.

Setting Up a Wargames Campaign

This is my 2^nd edition (1973) copy of, well, you can see for yourself:

I read this cover to cover (which sounds impressive but it only has 79 pages) and used its advice to set up my own wargames campaign.

It was a time-consuming task in the days when we didn’t have computers—I had to create individual slips of paper for every unit in the game, and there were hundreds of them. I worked on it on and off for several years, just about finishing it over the summer break before I went to university. This was in the belief that I’d be able to run it as a student, which unfortunately did not turn out to be the case. Maybe one day...

I mention this now to reiterate the point that I’d made many worlds before I began working on MUD. My wargames campaign was but one of them (albeit an alarmingly detailed one), my D&D campaign was another, Dr. Toddystone was a third, and there were others besides (at least one of which was even more detailed than any of those mentioned here).

If you want to make worlds, you can’t help it: you make worlds.

Gameaholics

Every once in a while a batch of near-identical articles shows up in my Internet vanity search. They’re on different sites and purport to be written by different people, but they’re basically all the same. They contain a particular phrase that makes them easily recognizable: “gameaholics named Richard Bartle and Roy Trubshaw”. There were close to 2,400 of these articles last time I looked (i.e., just now), but the number goes up and down. Sometimes it’s close to 5,000; sometimes it’s in the low hundreds.

This is the kind of phrase I remember, because Roy Trubshaw wasn’t and isn’t a gameaholic by any stretch of the imagination, so it’s basically bad journalism—or at least it’s based on what was originally bad journalism. The surrounding text changes a little: sometimes we’re “two” gameaholics, sometimes we’re “a couple of” gameaholics, sometimes we’re “a pair of” gameaholics, but the gist is always the same. These articles have been appearing for years—well over a decade.

It’s clear that they go through some kind of synonym-processing to make them look superficially different while actually being the same. One of the words in their dictionary has a mistaken form, too, as can be seen in another sentence in the articles (the one preceding the one that mentions Roy’s and my names). Some of the many variations of this include:

• “Whilst these types of games are widely well-known nowadays, they in fact have been in existence for quit some time”

• “Whilst these types of games are widely well-known nowadays, they actually have been in existence for quit some time”

• “While these kinds of online games are widely popular these days, they actually have been in existence for quit some time.”

• “While these online games are widely popular today, they in fact have been alive for several years”

• “Whilst these sorts of games have been at large obvious nowadays, they essentially have been in life for give up the little time”.

As you can see from the last one, this synonym system doesn’t always work. It would still have been wrong even if they’d spelled “quite” properly in the source document.

I guess the reason they do this is in order that the pages look valid to search engines so as to give the links on the page credence, the aim being to boost their profile. Whatever, it doesn’t work: as I said, I’ve been seeing these articles for well over 10 years, but only the newer ones show up in Google; presumably the earlier ones were detected by Google as being fraudulent and removed.

Every once in a while, though, along comes another batch.

Oh, for a Toolkit

The current vagaries of the MMO implementation process mean we have unitary, inflexible worlds that need to try to be as many things to as many people as they can be if they’re to recover their development costs.

If the effort involved in developing a 3D graphical MMO dropped to that involved in developing a text MUD (i.e., roughly on a par with writing a novel) then you could make your own. If you could build your own MMO as a hobbyist; you could charge whatever you liked for people to play; you could obliterate the accounts of miscreants on a whim; you could change the physics of the world to reflect your preferred playing style—in short, you could do what the MUD community has been doing for decades.

Does it really matter that only 1,000 people are ever going to see your MMO if it’s fun for you?

The Development Process

The development of regular computer games has three phases:

• Pre-production

• Production

• Roll-out

MMOs have a fourth phase:

• Operation

Pre-production is where the concept evaluation and project planning is done. It greatly reduces the risk of making mistakes in later phases, and therefore reduces the chance that Things Will Go Expensively Wrong. It’s the most exciting part of development (especially for designers), despite the lack of resources and the time pressure that traditionally accompany it.

Pre-production normally lasts no more than six months, although for very well-funded MMOs it could be longer (I know of one that took 18 months). It results in a set of deliverables in various states of completion. These will usually include a design document, a technical design review, an art bible, a production management assessment, assorted prototypes, plus sets of incoherent notes presented in smart binders.

The production phase is where the game is actually made. This takes around 18 months for a AAA console product, and up to three years for an MMO (although this is falling). Code is created (for the game itself and for development tools), artwork is executed (object models, texture maps, sundry images, whatever), level design is undertaken (dialogue written, behaviors scripted, 240 kinds of weapon specifications specified) and the result is something playable.

Roll-out consists of two testing phases. Alpha testing is a form of system testing undertaken in-house by the people making the game plus senior quality assurance personnel. Beta testing is performed by trusted people from outside the team, including hard-core players from related games.

MMOs have two types of beta tests. Closed beta comes first, and is invitation-only. This aims to test the MMO’s functionality and basic playability. It’s followed by open beta, which allows pretty well anyone in. This further tests playability, and allows for flood testing (i.e., can you really handle those 6,000 players per server that you think you can?); it’s also seen as a marketing tool, although this can easily backfire (there are many cases where MMOs had more beta-testers than they ever got post-launch).

The operation phase (or exploitation phase if you don’t mind the negative overtones) begins when people start being able to pay to play. Most MMO developers will take the original development team off the project at this point and switch to a live team (I’ve no idea why—developers such as Mythic and CCP, who don’t do this, have much better continuity and consistency). The live team’s duties include:

• Feature development

• Responding to acts of player cunning

• Network and server maintenance

• Customer and community support

• Keeping pace with technology (new hardware, new versions of middleware, and so on)

Regular AAA games are also increasingly getting an operation phase, to manage forum communities and check out user-created mods, but they’re tiny by MMO standards—just one or two people normally. In contrast, an MMO that has a development team of 30 at the end of its first year (5 designers, 10 programmers, and 15 artists) will probably begin the operation phase with a live team three or four times larger than that, mainly through the addition of 100 or more customer service representatives.

Not that the players will notice any of this, of course...

Change <player>

MUD1 had a spell that would change the sex of player characters; this enabled people to do both of the (what would now be called) quests that were gender-specific, instead of just one.

I deliberately made the spell easy to cast, so that people would be tempted to use it to annoy one another. Except, because there was no real difference between the sexes, I was using it to say there was nothing to be annoyed about. I put in the spell and the two quests entirely so as to make a political statement about gender.

As to what the politics of that statement were, well, if I could say it in words I wouldn’t have had to implement it in MUD1, would I?

What You Need to Know to Write an MMO (Today)

First you need to decide which kits, tools, libraries, and middleware you’re going to buy in to get the following functionality:

• 3D/graphics engine/renderer

• AI engine

• Animation package

• Art package

• Asset management software

• Audio package

• Back-end billing system

• Community management tools

• Compiler/development environment

• Database

• Load-balancing system

• Network library

• Object modeling/specification system

• Patching software

• Physics engine

• Project development tools

• Security system

• Vegetation-creation tools

• Web-creation tools

Having bought them, you and your team of programmers need to know how to sew the executable ones together with your own program code, then how to code the game mechanics on top of that, plus all the project-specific tools you require, while your vast army of artists are creating the graphics you’ve specified.

To do this, you must have knowledge of every API (application programming interface) for every kit, tool, library, and middleware you’re using.

If, when you create your virtual world, you don’t know the whole story from audio package right the way up to write-your-own-planet creation tools, you won’t be able to tweak every object, stretch every script, pack every data structure, or take every shortcut; less of your world will appear in the three-year window you get for its construction. You need to know all this if you’re going to do it.

Oh, you also need that little bit at the end where you design the game world itself.

Trains to the Wilderness

When you start off in a new MMO, it’s like arriving at a foreign railway station on a backpacking trip. With classes and “races,” the designers have provided trains that are guaranteed to go to interesting places. You want to shoot fireballs? Board the mage train! Quests are the engines that pull the carriages along.

However, trains run on rails. If you want to disembark somewhere other than at a station, well, you can’t. The design philosophy is all about controlling the player experience, and this same philosophy is applied for newbies and oldbies alike. It’s consistent—but players themselves aren’t consistent. Sure, when they started out they might have wanted to board the mage train, but if after watching the scenery for five months they wish to get out and explore the village of Melée Magic, they can’t. There isn’t a train from here that goes there.

Not all MMOs are like that, though. Some are more like hiking in the wilderness. There are few paths, and those there are often run out before reaching anywhere. You don’t know to start with whether you want to climb the mage mountain or cross the healing river or cut through the forest of swords. You might even want some of all three. You will, however, find a place that’s exactly right for you, eventually.

That’s eventually.

The MMOs that have this design philosophy are all about emergent content, but again they apply it consistently to newbies and oldbies alike. It’s great once you get into it, but you do have to get into it.

Ideally, MMOs should begin on rails, because most players really don’t know what’s ahead of them when they start out. They really do need guidance. However, after riding the train for a while they should be able to leave the rails and head off into the wilderness, because by then they have tastes too nuanced for crude character classes to capture.

In short, what we want here is for Alice and Dorothy to play together.

“I Want To Be a Lead Designer”

The lead designer is the person who is Keeper of the Vision^TM. Lead designers design the games.

They start off with the initial treatment (or vision document), and once it’s accepted by whoever green-lights the project they’ll develop this into a full-blown design document. They will usually keep notes to explain why they made their various decisions, but these are invariably incomprehensible to everyone, including the lead designer.

A lead designer may be the only designer for some games, but for MMOs they’ll usually head up a team of designers.

When lead designers come from a non-programming background (e.g., they have a history degree), they will consult with the project’s software planner (usually the lead programmer) to ensure that everything they want to have is programmable. The software planner will customarily lie in response, so as to make their programmers’ lives easier.

This may be why many designers do, in fact, have a programming background.

What You Need to Know to Write an MMO (Tomorrow)

That little bit at the end where you design the virtual world itself.

Floral Spaghetti Strap Dress

The Secret World has a cornucopia of art bugs. I particularly like this one:

You buy a sleeveless dress, you get a sleeveless dress.

DAMN YOU

As far as playing MMOs is concerned, I’m pretty well incognito. People who know who I am in real life can find out who I play as, and people who know who I play as could find out who I am in real life, but neither group has any particular incentive to do either.

Hmm, I’m sure some of those instances of “who” there should have been “whom”.

So, back in 2010, I was doing some wretched Valentine’s Day quest in WoW to get a pretty frock or something and I received a tell from one of the very few players who did know my name: “Did you see that?”.

“See what?” I replied.

“The shout!”

Okay, so I’d installed an add-on to filter out shouts and non-trade trade channel messages, in order not to have to subject myself to whatever conversations the resident 15-year-olds might wish to share with me; consequently, I hadn’t seen any shout. However, I checked the add-on’s equivalent of the Deleted Items folder and found this:

I’m almost certain that whoever it was who shouted those words had no idea that Richard Bartle was playing on that very server at that very moment, so it must have been that they were taking my name in vain because they actually felt I was in some small way responsible for whatever it was that was annoying them.

That’s kinda cool, actually! Except, of course, it should have been DAMN YOU RICHARD BARTLE AND ROY TRUBSHAW!!!

What You Need to Know About What You Needed to Know to Write an MMO (1978)

Only having read what I’ve written do I now realize how arcane what I wrote must seem to most of the readers of this book. It’s like something out of Finnegans Wake. Words with Chinese-whisper meanings drift gloriously between structural ones that convey form but not content.

In MMO terms, no one needs to know any of this, but someone needs to have known it. It’s a “deeper magic from before the dawn of time,” that informs all but has lost its meaning.

Rather like me, really.

So much knowledge, so little use.

DIY Options

If you’re thinking of making your own MMO, you basically have five options:

• Write everything from first principles.

• Use a commercial MMO engine.

• Use an MMO software development kit.

• Use an integrated creation and hosting solution.

• Embed it within a host virtual world.

Writing from first principles gives the most control, but involves a great deal of effort, which costs a great deal of money. It is in attempting to avoid this expense that the other options have arisen.

We got all of these options with textual worlds, but with graphical worlds only the first has resulted in commercial success. The rest have been tried, but none have yet taken off.

So, there’s a bit of will-the-future-repeat-the-past? for you to ponder on.

Alice and Dorothy Play Together

The central design philosophy of Dorothy worlds is one of direction: Dorothy is wary of the new world she finds herself in and wants a path she can follow to get through it. She’s about structure. She represents modern game worlds epitomized by World of Warcraft.

The central design philosophy of Alice worlds is one of being: Alice finds merely existing in the new world interesting. She’s about freedom. She represents old-style balanced worlds such as MUD1.

The central design philosophy of Wendy worlds is one of expression: Wendy wants to live in the personal fantasy that she has created. She’s about statement. She represents the modern social worlds such as Second Life.

Because of the Great Schism, Wendy worlds define themselves as not being Alice or Dorothy worlds. Dorothy worlds define themselves as not being Alice or Wendy worlds. Alice worlds look on both as estranged children. History has locked Wendy and Dorothy worlds into design philosophies that concern what these worlds aren’t almost as much as what they are.

Alice worlds are newbie-unfriendly but provide the depth and freedom that oldbies crave. Dorothy worlds are very newbie-friendly but oldbies who don’t want their hands held feel disenchanted. So, why not start off as a Dorothy world and then switch to Alice for the elder game? These decades-old philosophical differences have to persist no longer!

How this would work: you’d start off using a character pack optimized for one kind of experience (say, mage). You can diverge from it any time you like, but are given advice so you can’t do it without knowing what you’re letting yourself in for. Quests will initially be hand-crafted, but the more you play then the more will emerge from player interactions (such as putting want ads in an auction house, but potentially much more complex than this). Eventually, you’ll segue into a freeform game.

Alice and Dorothy can play together because they both have the conceit that they’re separated from Reality. This is possible because both are in worlds not of their own making. Sadly, Wendy is in a world of her own making. She can build Alice and Dorothy their own Wendy houses to play in, but she can’t play with them. They can still be friends, of course, they just can’t play together.

Today’s MMOs are victims of their own orthodoxies. They operate within artificial boundaries for obscure historical reasons few people are aware of any more. Players and designers sense there’s something wrong, but not quite what. Fortunately, by understanding the cause, the effect is easily removed.

As a final word of caution here, I’m not saying that how things once were means that’s how they should be. History can inform the present and the future, but it shouldn’t dictate it. Times change.

Nevertheless, Alice and Dorothy can play together, and we’d get better MMOs if they did.

“I Want To Be a Level Designer”

No, you actually want to be a lead designer; you’re only a level designer because that’s one of the steps on the ladder.

Level designers create episodic gameplay experiences within the framework of the overall design. The name “level designer” comes from the practice of assigning the creation of self-contained game levels in early games to junior designers. Nowadays, it just means a junior designer (but is still hanging on as an authentic job title).

Level designers often work with tools created by the programming team specifically for them. For example, a level designer for an MMO might have the job of creating monsters. Said designer will find out from the senior designers (the content team—that’s content as in “contained by,” not as in “at peace with the world”) what monsters are allowed that will fit the fiction and animation, then having chosen one he or she will design its properties and behaviors by filling in fields in a design tool’s window.

Level designers can also write a lot of scripts. This is another reason why it’s useful for designers to have programming experience.

“I Want To Be a Lead Programmer”

Why?

The lead programmer (who will often double as the software planner and the lead architect) is basically responsible for the integrity of the project’s software. They monitor the other programmers to make sure they’re doing what they’re supposed to be doing and are on schedule. If they find any problems, they’ll set the producer on the offenders.

Almost always, the lead programmer will be the most technically able programmer on the team. The role therefore carries much kudos among programmers. However, because they have to manage the other programmers, lead programmers typically only spend half to three quarters of their time actually programming, which is a clear waste of their skills that developers nevertheless seem happy to accept.

Given that programmers program because they like programming, though, why would they want to be a lead programmer if that means giving up time they could spend programming?

Hmm, well, they are paid more than regular programmers.

Plodding

In the text MUD days, we regarded time as a poor substitute for skill. If you could get to the top through time alone, with little or no skill, we called it plodding; it was regarded as A Bad Thing.

Today’s MMOs actually protect plodding as a leveling strategy. Spend enough time trying to get to the level cap and you will indeed get to the level cap. There’s nothing to stop you. Worse, you actually need to spend the time to get there, rather than blitzing your way through because you’re skillful.

The fact that plodders outnumber non-plodders probably explains this.

Back from the Mists of Time

Every so often, I receive invitations to attend industry conferences (“a whole day for only £816.63!”) sent by people whose actual understanding of the industry is so limited that they don’t know why they shouldn’t be sending such invitations to me. These conferences invariably include perennially favorite topics such as “the importance of community,” “multi-player management,” and “business models” that were old when I went to the predecessors of these conferences in 1995.

I don’t know whether this is a good thing or a bad thing. I don’t know whether to be amused or disgruntled.

I must be getting old when not only do I have a sense of déjà vu, but I have a particular sense of déjà vu that I feel I’ve had once before.

Wow!

Look at the size of this Defias Thug’s weapon!

World of Warcraft’s collision detection is deliberately turned off for interactions between pretty well anything that moves. This saves on all those tedious calculations, but it can lead to unfortunate comedic experiences.

Criteria for a Name

On the assumption that we won’t be able to call MMOs MMOs for much longer, and that the umbrella term virtual worlds has already almost slipped from our grasp, how should we go about creating a new term that is built to last?

What we need from such a term is:

• Pronounceability: No acronyms you have to spell out.

• Unclaimed: Someone already staked out Virtual Reality. The acronym for Virtual Worlds is used by Volkswagen.

• Inclusive: If it mentions games, the Second Life crowd won’t like it; if it mentions user-created content, the World of Warcraft crowd won’t like it.

• Exclusive: Bingo is a “massively multiplayer online game” when you break the terms down and see if they apply.

• Future-proof: Does it fall prey to a distinction between free-to-play and pay-to-play worlds, say? Would it still work if stereoscopic 3D head-mounted graphics became nearly universal?

• Neutral: Sorry, Persistent Environment Games, but your acronym is too smutty.

• Descriptive: It should at least hint as to what it refers. If only “Those Online Things” were acceptable.

This is why I expect to go to my grave without there being a commonly-accepted and long-lasting term for (what we currently call) MMOs.

What’s in a Name?

Most people probably aren’t aware that memory sticks are so called because that’s what Sony called its flash memory product in 1998. However, if you were to set up as a manufacturer of flash memory, it’s something you really ought to know. You wouldn’t even consider calling your product “memory stick”.

July 28th, 2015: Pure Bang Games get their project greenlit on Steam. It’s “an open-world exploration and survival game, inspired by old school M.U.D.s and Minecraft.” Their game is called MUD.

That’s the trouble with history; it’s all in the past.

“I Want To Be a Lead Artist”

The lead artist is (as I’m sure will not come as a surprise) the person in charge of ensuring that all artists working on a project are producing output that is consistent with the Vision, and with the work of each other.

As with lead programmers, this means that the best artist has to give up time doing what they do well (in this case, art) to perform a task to which they may be ill-suited (monitoring other people). However, lead artists only tend to spend maybe 10%-20% of their time looking at what the art team produces, because it’s so much easier to see flaws with art—poor work is obvious at a glance.

The lead artist will usually be the person who creates the look of the game and the rules that the art team will follow (compiled as the art bible). This will have been done in consultation mainly with the lead designer, but also (for animation) the lead programmer. MMOs have a long shelf life, which can mean state-of-the-art, er, art at launch looks old and jaded five years later; art bibles should therefore be strong enough to be dusted off and re-used by possibly a new team several years away.

For a large art team, there may be sub-leads with responsibility for specific areas of artwork—static or animated, for example.

Large developers sometimes pool their artists, so any individual artist could be working on several different projects at once. This is because the requirement for artwork can come in ebbs and flows. For consistency, lead artists will remain with one project, however. Needless to say, this can lead to strife when two leads from different projects tell the same artist that they want a ton of work produced RIGHT NOW.

Artists in general are renowned for not getting along with technology and for being independent free-thinkers who don’t take well to being bossed around. Artists who work on computer games are no exception. If you find a project for which all artwork is reliably logged in the asset management system, watch out: it means the producer has a taser hidden somewhere.

“I Want To Be a CS Lead”

Customer service is an important part of MMO operation (not that this is reflected in the salaries of those working in it). Also known as community management, it’s basically the player-facing part of an MMO development team, handling pretty well anything that involves speaking to the customer base. This includes fielding general queries, fire-fighting forum flare-ups, managing the web site, judging competitions, banning people for misbehaving, dealing with billing enquiries, preliminary bug investigating, reporting to the appropriate local agency people threatening suicide, telling players what they would have known if only they’d read the manual, and so on and so forth.

CS representatives (or CSRs) are, in common with quality assurance technicians, prone to burning out. Their job is dealing with problems that to those concerned are very often regarded as emergencies. As a result, most players they deal with are uptight and pushing for answers or action, which leads to a lot of pressure.

The CS lead has the doubly tricky job of training and managing the CSRs while fending off the rest of the company’s attempts to talk to players directly. One misplaced word from a programmer can lead to the dissipation of months of carefully-managed player expectations. One premature press release by the marketing section can cause a storm surge on the forums as people weigh in with awkward questions and indignant complaints.

Good CS leads are hard to find, but they do exist, and they do emerge from the pool of CSRs. It’s the kind of job which can often be very rewarding (people are genuinely happy if you solve their problem for them), but also a little frustrating (“oh look, another death threat made it through my spam filter”).

In the early days, when MMOs (in the form of text MUDs) were relatively small, they were usually managed by senior players answering to the designer/programmer. As such, I personally acted as CS lead for MUD1/MUD2 for over 20 years. This experience means I’m surprisingly good at it, but I have to say I don’t like doing it at all—it’s hassle, hassle, hassle, all the way. CS leads, who have chosen this as their career path, have my undying admiration. I know how tough their job is.

Engines of Creation

Most independent MMO developers don’t use a commercial MMO engine, they write their own. Okay, so they may use Unity or something for the graphics, but the rest of the system will be bespoke. The reason for this is that commercial engines are too expensive, they don’t do what you want, and you can write something better anyway (“and sell it”).

This situation won’t last, as off-the-shelf engines are becoming more and more powerful and attractive (that is, free). It is how things are right this moment, though.

Now although indie developers may not like the extra work, it turns out that programming a hand-crafted engine is exceptionally good for creativity. The more you write in-house, the more different your world will be to everyone else’s—there’s a much greater possibility for divergence. This is why there was such a variety of worlds at Adventure ’89.

It won’t be like this for ever, though. With textual worlds, it went:

• 1989 TinyMUD

• 1989 LPMUD

• 1990 DikuMUD

• 1990 MOO

• 1990 TinyMUCK

• 1991-present One of the above

The capacity for innovation is reduced once game engines become a viable option. People take the short cut then have to rely on features that if they’d implemented them on their own would be entirely different. The physics, the combat system, the economy—you get the idea.

Any new medium only gets one window of exuberant creative expression. For graphical worlds, we’re beginning it right now.

At least, I hope we are. If we’ve already been into it and come out the other side, we’ve got depressingly little to show for it.

Benefits of a Federation

Here’s a set of notes I composed with regard to setting up a federation of MMO developers. My guess is that I wrote it around 1987:

• BENEFITS OF A FEDERATION

• (IN NO PARTICULAR ORDER!)

• Consistency across games for similar commands.

• Guardian of standards in the face of PD [public domain, i.e., free] MUDs.

• Coordination of press releases.

• Organize conventions.

• Joint stands at shows.

• A common method of rating games for sex/violence, like with films.

• Better we do it than amateurs!

• Founder members get a say in organizing it, rather than taking what comes in the inevitable federation that will arise instead.

• Stand up to large competitors (e.g., Fujitsu).

• Stand up to large companies (e.g., BT, CompuServe).

• Common legal contracts to use, e.g., non-disclosure agreements, royalty agreements.

• Central mailing list.

• Blacklist (optional) for eternal troublemakers.

• Joint protocols for graphics, etc., so can share FEs [front ends, i.e., client software].

• Focal point for media interest in these games in general.

• Code of conduct, to protect us from “on-line porn available” charges in NOTW [News of the World, a now-defunct sensationalist newspaper].

• Guidelines for charging strategies.

• Newsletter.

• Pressure group v e.g., BT, government.

• (Members would be commercial systems, affiliates for amateurs)

• “Commercial” needs a definition!

• Companies are members, not individuals?

• Example: Government decides to ban MUAs because “they’re the same thing as chatlines”. What can we do unless we’ve organized beforehand?

• Contact list for e.g., U.S. companies

• Name? Awkward, because there is no common term for these games.

• Soc. of Multi-User Adv. games—SMAUG

• Society? Federation? Organization? Foundation? Association?

• Independent?

• International?

• The games themselves—Multi-User Adventure Games?

• Multi-/Many- User/Player

• Adventures?

• Role-playing?

• Text-based?

• Virtual/Alternate realities?

Entirely unsurprisingly, we’re no closer to the “inevitable” federation now than we were then.

Falling Proud

A question I’m something asked by interviewers is “Which of MUD’s achievements are you most proud of?”.

Why would I be proud of any of them?

It’s all lost opportunities from my point of view. I see unreached potential, I see glorious possibilities, I see liberty, I see the ability to BE! Then I look at what we have, and I despair.

The gifts that MUD held forth and promised, humanity will eventually obtain; however, it will be from a reinvention of the concept, not a rediscovery.

I’m not proud of MUD’s achievements, I’m frustrated by them.

Reading Matter

It used to be that I’d read every academic paper about MMOs because I’d written them.

Then, other people started writing on the subject. I eagerly tracked down and read every single piece of work I could find. Even in the late 1990s, it was possible to do this: a newcomer to the field could sit down for three or four weeks and read in full pretty well every major article on the topic.

Well, in English, anyway—I had some foreign-language papers too but I couldn’t make head nor tail of those.

MMO books started to come out. I bought and read each and every one I could find. Every couple of months, I would spend a few hours in the largest London bookshops looking through the indexes of anything promising, to see if there was a mention of MUDs, MOOs, or (sometimes) me.

By the time my book Designing Virtual Worlds came out in 2003, things were starting to change. I could keep up with the books, but I had a backlog of papers starting to build up, mainly because of a few lengthy MA and Ph.D. theses that took ages to read.

Shortly after, I started to discover areas of research that, worryingly, I hadn’t come across before—whole seams of it, waiting to be mined. The reason I hadn’t heard of it earlier was because the researchers involved hadn’t heard about MMOs. They had their own names for what at the time we were calling “virtual worlds,” and were put out that no one would adopt their terminology.

In about 2005, the floodgates started to open. Partly because academics were writing more papers, but mainly because Second Life and World of Warcraft garnered so much attention, papers started to appear faster than I could read them. Some areas, such as Serious Games, really rocketed. The quality of the papers went down, though; I was reading material that treated what we’d known for years as if it were a new discovery, or that made big claims on flimsy evidence, or that had poor scholarship, or that misinterpreted facts, or that was just plain wrong. My pile of papers to read got higher and higher. I started to get a pile of books, too.

Still, I kept collecting every article on virtual worlds that I could find. If I couldn’t read it right away, I might get time later. Often, I did get time later. Whenever I came across a bunch of papers, I would print them off and put them in my pile to read. The pile just seemed to get taller and taller, though.

On August 26th, 2010, I received a link to a journal in my e-mail. It wasn’t anything out of the ordinary—Learning, Media and Technology vol. 35(2)—and it included some papers that I did actually want to read (and indeed did go on to read). However, through no fault of its own, it was the straw the broke the camel’s back. I lost the will to print off every single paper in it to add to my repository. Merely being about virtual worlds wasn’t enough of a reason to collect papers any more. There were just too many.

I do try to get hold of books still, but not for archival purposes; I get them because I want to read them. Twenty years from now, there’ll be no need for someone to have collected any of them, because they’ll all have been digitized (and, most probably, neglected). A couple of hundred years from now, historians will simply search for articles (“sources”) that support whatever point they are trying to make; they won’t know the good papers from the bad—and they won’t necessarily care, either, so long as what they find substantiates their argument.

Still, I wasn’t ever collecting papers for posterity, I was collecting them to read. It took a while, but I finally had to recognize what had been obvious for some time. I don’t have the time to read everything, and a good deal of what’s out there isn’t worth reading anyway. I have to be selective instead.

Besides, I’ve run out of shelf space in my office at the university.

As You Are Now, So Once Was I

Of course things have changed since virtual worlds went graphical. The main change isn’t so much the graphics itself as the fact that there are orders of magnitudes more players in single incarnations of MMOs as a result of the introduction of said graphics.

As I’ve suggested before, the difference between textual and graphical virtual worlds is like the difference between silent movies and talkies. Once sound came out, it killed the talkies pretty well stone dead. That didn’t mean knowledge gained in making silent movies was suddenly inapplicable “because some people like movie stars for the way they talk”. Neither did it mean that people who grew up making silent movies were incapable of accepting progress.

There is exciting and innovative work going on in textual worlds, but this doesn’t alter the fact that they’re a dated form. You can legitimately refer to them as old hat because frankly, yes, the concept itself is old hat. You would be wrong, however, to dismiss as flotsam what has been discovered through these worlds.

Look into the future when the virtual world equivalent of color in the movies comes along—virtual reality hardware, say, or voice synthesis, or believable artificial intelligence. You, an expert in black-and-white talkie worlds, may be very keen to encourage this exciting phenomenon—imagine the possibilities! It’s going to come as a bit of a jolt, therefore, when you discover that people new to the field start dismissing as evolved-away all the work that you and others have done on black-and-white MMOs, not even caring to look at it before making their judgment. Indeed, they may want not to look at it.

So, if you want to ignore the past, that’s your right. However, bear in mind that the present will, in time, itself become the past. If you can’t believe that all your knowledge of MMOs will be irrelevant to the MMOs of the future, at least afford the same acknowledgement to the players and designers of the MMOs that came before yours.

“I Want To Be a QA Lead”

Quality assurance (known universally as QA—so much so that there are probably people working in QA right now who don’t know what the acronym expands to) is basically about testing. The QA lead supervises the people who do this testing, telling them what to test.

There are two main types of testing that need to be done: technical and gameplay. Technical testing is done by QA technicians, and is focused on the software. All the code paths, no matter how trivial, are investigated to see whether they do what they’re supposed to do (nothing more, nothing less). Gameplay testing is done by playtesters: they check that the game plays well, that it has no exploits, and that it’s balanced. They don’t check that it’s fun, because nothing is fun if you spend eight hours a day doing it for months on end.

“QA technician” and “playtester,” while formally separate roles, are often used synonymously. It depends on whether you want to give the impression of respecting those who do it or of disrespecting them.

The QA lead, in collaboration with the producer, determines what needs to be tested. It’s usually better to leave the other leads (designer, programmer, and artist) out of it, because it’s their work that’s being tested. Oh, and yes, artwork and modeling is tested: if you like the idea of walking your avatar into every single wall in an MMO to make sure it functions as a wall, QA is your field.

The QA lead draws up the test plans and assigns different areas to different QA technicians. QA technicians are not, however, easy to manage, because most are not there out of a love for testing: they either want to get a foot in the door as a designer, or they have the naïve idea that playing games for a living would be paradise. It’s also possible that they want your job, in order to get the producer’s. Most QA technicians burn out within two years, but some few do find they have a genuine aptitude for the work. These are the ones who will go on to be QA leads.

QA has the lowest-paid positions in game development, but a good QA lead is so hard to find, and the consequences of bad QA are so expensive, that talented people are rewarded well.

Personally, I’d be burnt out in under a month.

The Personality of a World

One of the legacies of MUD that has reached today’s MMOs is its playful spirit. I don’t know whether today’s designers generally consider that their MMOs have a personality, but I certainly did for MUD and designed it accordingly.

I used to run my own postal games fanzine in my teens (called Sauce of the Nile) (I did say it was in my teens). People would subscribe to several such fanzines, and I realized that readers interacted with each one differently by projecting personalities onto them. This was mainly because Sauce of the Nile had a number of subzines that had different personalities, making the clashing contrast rather irritatingly obvious.

I wrote MUD with a single voice in mind, in the hope that players would find it amusing, engaging, affable, and likeable. I wanted them to feel that it—the virtual world—was in some sense a friend. I say “voice” here, but the language it spoke in wasn’t English (although as a textual world, obviously it used English); the language it spoke in was the make-up of the virtual world itself. If I wanted the players to like the world, then the world was my medium; the text was merely a means of bringing MUD’s world to life, as graphics are today. Playing MUD was meant to be a playful activity, so the world itself had to be playful.

In the main, today’s MMOs still have that aesthetic. I’d like to think it’s because the designers of today were the players of yesterday, and they picked up on the idea either consciously or subconsciously and went with it after some thought. However, I suspect that in the majority of cases, few designers have given the matter much consideration and have just carried on with what they know works as part of the paradigm.

Nevertheless, I still have the vague feeling that whenever anyone sits down to play an MMO, in some tenuous sense I’m there playing with them.

“I Want To Be a Producer”

A producer is the project manager for a development team. Indeed, “project manager” may be their job title.

The producer plans the overall development schedule and keeps everything on track. They badger leads when their groups fall behind, and they badger management when the leads need more resources (which is not quite the same as when the leads say they need more resources).

The producer’s job is more onerous than this sounds, to the extent that on big teams they may have one or more assistant producers to help them.

As you can no doubt deduce, producers annoy management by asking for more money/people/time to complete the project, and they annoy developers when management says they can’t have it. Thus, they are hated by all. This perhaps in part explains why they get paid more than anyone else in the core team except the elitest of programmers.

Design by Committee

There is reluctance among some game developers to accept design as a skill. I have in my time come across several companies that design games by committee. They get everyone in the company together and brainstorm ideas. It’s like a badge of pride for them that “even the receptionist” helps design games.

No!

No one writes a novel, paints a picture, or composes a symphony by committee. Well, they do, but the results suck. It’s the same for games.

Most of the companies that practice design by committee seem to make serious games, by the way, rather than commercial games. That probably explains it.