Keywords

1 Introduction

Software Process Improvement (SPI) has contributed significantly to the improvement of development and other activities related to software. Consequently, new concepts and technologies are eventually consolidated to challenge and expand SPI and impact its practical and research context. The publication of the ISO/IEC 15504 standard set in the early 2000s, for example, caused this impact and expanded SPI by consolidating concepts, requirements and examples of reference models and process assessments methods.

Recently, the requirements for process measurement frameworks defined in ISO/IEC 33003 [3] introduced relevance challenges and changes in SPI research and practical context. ISO/IEC 33003 states that a “process measurement framework shall identify and address a single process quality characteristic, which shall be defined on the basis of a multidimensional construct and as a set of process attributes” [3]. Before the publication of the first set of ISO/IEC 330xx family of standards [1,2,3,4,5], SPI was strongly tied with process capability. Therefore reference models, assessment methods, and improvements methods for SPI are tied with process capability concept. Process measurement frameworks, based on process quality characteristics other than process capability, such as, for example, safety, agility, and systems thinking, expand SPI to broaden its application.

In addition, distinct process practices of generic profile group for Very Small Entities (VSE) from ISO/IEC 29110 [6, 7] and distinct specific practices for each capability level of each practice area in the recent version of CMMI (CMMI Development V2.0) [8], to characterize process evolution towards capability, instead of generic capability practices, also caused challenges and changes in SPI context. There are additional challenges and changes in the SPI context that are identified later in this article.

Hence, comprehensive methodologies for SPI, such as, for example, PRO2PI Methodology (Process Capability/Modeling Profile for Process Improvement) [9,10,11,12], should be analyzed and evolved to consider this challenging new SPI context. Such analyses and evolution are relevant SPI research and practical theme.

Therefore, the main objective of the research presented in this article is to analyze PRO2PI Methodology in face of these recent challenges and changes in SPI context and then design PRO2PI evolution to guide better process improvement in this context. To guide the achievement of this objective, it is decomposed into four guided-oriented sub-objectives:

  1. (a)

    Identify recent challenges and changes in SPI research and practical context;

  2. (b)

    Describe current utilization of PRO2PI;

  3. (c)

    Analyze PRO2PI in face of its current utilization and identified recent challenges and changes in SPI context; and

  4. (d)

    Design PRO2PI evolutions to consider those recent challenges and changes in SPI context.

The meaning of SPI needs some considerations. First, the term software in SPI is used here for historical reasons. It actually means a broader scope, including systems, services and others. Actually, SPI should be knowledge working process improvement, including software, systems, services and other processes. There is a shift to knowledge working intense organizations, as identified by Peter Drucker [13]. What has been done for software process, including SPI, can be used for knowledge working processes. Second, although ISO/IEC 330xx Family of Standards, and its previous versions as ISO/IEC 15504 Set of Standards [14, 15] are international standards of process assessment, they have been causing an impact in SPI as well. Therefore they should be considered as reference for SPI. ISO/IEC TR 33014:2013 [16], for example, provides a guide for process improvement. Therefore SPI is considered here as knowledge working process improvement, which includes assessment based process improvement on software process.

Following the four guided-oriented sub-objectives defined, this article is structured into six sections. This first section is an introduction. Second section presents the history and current version of PRO2PI Methodology. Third section introduces two recent utilization of PRO2PI. Fourth section presents six identified recent challenges and changes in SPI research and practical context, and comments on their impact on PRO2PI Methodology evolution. Finally, fifth section presents some conclusions.

2 PRO2PI Methodology

PRO2PI is an innovative process improvement methodology. It evolves current model based SPI towards a modeling driven SPI. PRO2PI development began in 2002 based on my experience in process improvement with ISO/IEC 15504-5 Process Assessment Model [9, 10]. The motivation for improvement was the techniques to choose the processes to be used in each improvement cycle in each company. The processes chosen for improvement and their respective current and intended capability levels formed the Process Capability Profile for Process Improvement. A profile was chosen in each company based on the analysis of its characteristics, problems and objectives. A profile is validated in terms of the degrees of attendance to eight properties (quality characteristics): relevant, feasible, opportunistic, systemic, representative, traceable, specific, and dynamic.

The PRO2PI Methodology has been conceived in many cycles of exploration, application and consolidation following the industry-as-laboratory approach proposed by Potts [17]. Potts argues that the traditional research-then-transfer approach has problems because it treats research and its application by industry as separate, sequential activities. In the proposed research approach, there is stronger connection at start because knowledge of problem is acquired from practitioners in industry.

During these cycles, we identified two other types of profiles that, in addition to Process Capability Profile, also drive a process improvement cycle. We named them Process Enactment Description and Process Performance Indicator. A Process Enactment Description is structured with life cycle, roles, activities and artifacts. A Process Performance Indicator model is structured with information needs, information product, indicator and measures.

These three types of profiles were identified using a modeling view of SPI [18]. They are actually different models of the process. During a cycle of SPI, these three types of process models should drive the process improvement (Fig. 1).

Fig. 1.
figure 1

Process as a model and types of process models for SPI [18]

Full size image

In a previous article [18], I presented a modeling view of SPI. First I presented a set of related concepts from Bézivin [29], Favre [28] and other authors [30], including concepts of model, its three elements (system, intended goal and aspects), how it follows the Limited Substitutability Principle, metamodel (as a model of a language of models, rather than a model of a model), how a model is used as a specification model or as a descriptive model and theirs co-evolution, modeling and chain of models. Then I concluded “modeling is essential to software process improvement, because every human action is preceded by the construct (implicit or explicit) of chains of specification and descriptive models”.

Then I proposed chains of models to represent process and three most relevant types of process models for SPI: Process Capability Profile, Process Enactment Description and Process Performance Indicator. The relationship between a model and a system is “isRepresentedBy”. After this modeling view, the methodology evolved from “Process Capability Profile” to “Process Modeling Profile”.

PRO2PI is named as a methodology in the meaning of methodology used by Schreiber and Akkermans [19] in Knowledge Engineering. A methodology is a sequence of feedbacks cycles (as a pyramid) with a worldview based on a set of principles that form the baseline of a methodology. This worldview is grounded in theories that provide the essential concepts for establishing the methodology. The methods and tools provide the key to enable the practical application of the methodology. The use of this methodology produces feedback that feeds the other “layers” of the pyramid and enables the evolution of the methodology.

PRO2PI Methodology has been developed with five major methodological components (Fig. 2). A definition phrase, initials and “label name” identify each methodological component of PRO2PI. The “label name” is a name of a Brazilian music album with a cover that resembles in a specific manner the component. The methodology diagram in Fig. 2 also has a “label name”, in this case, “Maritmo”.

Fig. 2.
figure 2

Methodological Components of PRO2PI Methodology [18]

Full size image

The five methodological components are:

  1. (a)

    Process Modeling Profile Metamodel (PRO2PI-MMOD, MMOD or Geraes): A metamodel of different architectures of good practices reference models to allow that elements from them could be used in a Process Modeling Profile to guide an SPI Cycle. Currently, there is a metamodel for Process Capability Profile. This metamodel defines a consensual agreement on how elements of a process should be selected to produce a given PRO2PI as a model of a process.

  2. (b)

    Process Modeling Profile Quality Model (PRO2PI-QMOD, QMOD or Passarin): A quality model of a Process Modeling Profile. This model is used to verify the quality of PRO2PI to guide an improvement cycle. Currently there is a quality model for Process Capability Profile.

  3. (c)

    Method Framework for Engineering Models (PRO2PI-MFMOD, MFMOD or Livro): A method framework to develop a reference model based on context and characteristics of a segment or domain.

  4. (d)

    Method for Process Improvement Cycle (PRO2PI-CYCLE, CYCLE or Uakti): A method to guide a process improvement cycle driven by a dynamic PRO2PI. This method includes a function to define, update and use a PRO2PI. It also defines six phases. The first phase Prepare for improvement cycle starts after a decision and commitment for improvement. The second phase is Establish improvement references. The third phase is Prepare for improvement actions. The fourth phase is Implement improvement actions. The fifth phase is Prepare improvement institutionalization. The sixth phase is Institutionalize improvements.

  5. (e)

    Method for Workshop to establish a Process Modeling Profile (PRO2PI-WORK, WORK or Mapa): A method to guide the implementation of the first four phases of PRO2PI-CYCLE in a VSE. This method has been developed to be used in the first phases of traditional process improvement cycle or PRO2PI-CYCLE cycle.

Recently, a new methodological component has been added. With the increase in its use and the importance of education of SPI, PRO2PI-WORK4E (and its evaluation model) consolidated as the sixth element. PRO2PI-WORK4E (“for (four 4) Education”) (or simply WORK4E) is a customized version of PRO2PI-WORK method to teach SPI. Both WORK and WORK4E guide the first phases of an SPI cycle and the learning process of SPI. While WORK focuses on starting an SPI cycle with learning SPI (“doing SPI with learning SPI”), WORK4E focuses on learning SPI by starting an SPI cycle (“learning SPI by doing SPI”) [20].

3 Recent Use of PRO2PI: WORK4E and CERTICS

In the last four years, from April 2014 to May 2018, I have been working with PRO2PI methodology in two components. First, I am reviewing, using and evaluate WORK4E. Second, I am using MFMOD for the development and evolution of CERTICS Reference Model as a model for innovation in software.

Recently, in 2016, 2017 and 2018, WORK4E has been used in three editions of a course to introduce SPI. An evaluation of 2017 edition was performed with a specific evaluation model [20]. Here, I presented an analysis of the results from last three editions [2016–2018] in terms of the PRO2PI Methodology. Such analyses are not described in the previous article. Here we concentrate on the SPI cycle developed by the students in terms of PRO2PI Methodology.

The courses are for professionals, working in IT, with a degree in IT related area. The duration is short, around 40 h, distributed in 1 to 2 months. The method guides courses where the possibilities for the production or construction of SPI knowledge are created with the development of a proposal for an SPI cycle in the professional’s work environment (“learning SPI by doing SPI”). Each group of students develops a proposal.

The key activity from each group starts with the identification and description of an Organizational Unit, problems and objective and goals for a process improvement cycle. Then each group analyses five to six processes, processes areas or practice areas (named as process area in the course) in terms of importance and risk. Then they estimate process capability level of each one. Finally, each group selects two process areas to be the PRO2PI.

Figure 3 presents a representation of each process area presented, with its identification, name and source reference model, and a version of the choices of two of them by each group to be a PRO2PI. These choices were in 2016, 2017 and 2018 editions of an introduction to SPI course with PRO2PI-WORK4E. Each line connecting two process areas indicates them as process areas chosen to be a PRO2PI.

Fig. 3.
figure 3

Twenty PRO2PI with two process areas each

Full size image

In 2016 edition, five processes area were presented: PP, SD, CRM, TEC and DE. There were five groups. Three groups chose PP and CRM as the process area and process to be a PRO2PI. Other group chose SD and CRM. Another group chose PP and TEC. In 2017 edition, five processes area were presented: PP, SD, DRE, TEC and DE. So, for this edition, DRE process area replaced CRM. There were seven groups. Two of them chose SD and DE. The other five groups chose PP and SD; PP and DRE; PP and TEC; and DRE and TEC. In 2018 edition, five processes area were presented: PM, TS, NEC, TEC and DE. So, for this edition, PM, TS and NEC replaced PP, SD and DRE. There were eight groups. Two groups chose PM and TS. Other two groups chose PM and TEC. The other four groups chose PM and NEC; PM and TEC; TS and NEC; and TEC and DE.

As there were a diversity of characteristics of the twenty Organizational Units and also a diversity of their problems and improvement objectives and goals, a diversity of selected pairs of process areas were expected. This diversity was achieved. In general, the argument, expressed in each one of the twenty proposals for improvement, for choosing the pair of areas was considered satisfactory and consistent with the Organizational Unit, its problems and its improvement objective and goals.

The other recent utilization of PRO2PI is related with the CERTICS Model. The Brazilian Government established a public policy instrument to identify and stimulate software production resulting from technology development and innovation carried out in Brazil. In order to accomplish this effort, CERTICS software process assessment methodology was created and established in Brazil. CERTICS was developed using PRO2PI-MFMOD. Its construction has been based on the reality of local software development organizations, in effort to achieve consensus within the community of interest, and guided by methodological references including the ISO/IEC 15504 (SPICE) Standard. CERTICS Methodology includes an Assessment Reference Model and an Assessment Method [21].

CERTICS Assessment Reference Model is a model for innovation and defines four competence areas (following the concept of process of a Process Assessment Model – PAM) named as Technological Development (DES), Technological Management (TEC), Business Management (GNE) and Continuous Improvement (MEC). Each Competence Area involves, with different emphases, aspects of both technological and correlated competencies. Each Competence Area is characterized by a key question, followed by a brief description and a set of Outcomes. Each Competence Area is achieved if their outcomes are achieved.

CERTICS Assessment Reference Model defines only capability level 1 for each competence area, and maturity level 1 with the four competence areas as the Basic Process Set with no additional elements. CERTICS Assessment Reference Model is compliant with ISO/IEC 15504 Requirements for Process Reference Models, Process Assessment Models and Organizational Maturity Models [21].

CERTICS was developed to be used for certification. Therefore it defines only maturity level 1. In order to be used as improvement model and to facilitate its usage with other models, three evolutions are performed, using parts of PRO2PI-MFMOD. First, we knew that the granularity was too high, compared with other models. So, instead of relate an area as a process, we relate each area as a group and then an outcome of an area as a process. Each relevant element of a CERTICS outcome became an outcome. Second, we rewrote sentences from passive to active voice. Finally we included the capability dimensions, for levels 2 and 3, following ISO/IEC 33020. A prototype version has been developed and used in two editions of a CERTICS Process Assessment course. This prototype version is named CERTICS-SAE Process Assessment Model. SAE stands for Simplified, Adapted and Extended. It is simplified because it is composed of only five of sixteen processes (identified by DES.1, TEC.1, TEC.3, GNE.3 and MEC.1) of the Model (PRM), which are detailed with Basic Practices, and four Capability Levels (0, 1, 2 and 3), of which three are detailed with Process Attributes (Table 1).

Table 1. Summary of the CERTICS-SAE process assessment model
Full size table

CERTICS-SAE was used in two editions of two classes on process assessment model and process assessment, in 2016 and 2017. In those classes, the original CERTICS and the CERTICS-SAE models were presented. CERTICS-SAE was presented as an example of model which a typical structure. Process assessment was presented and exercised in practical cases with CERTICS-SAE.

4 Recent Changes in the SPI Context and PRO2PI Methodology

In my view, there are six recent challenges and changes in SPI research and practical context more relevant for a process improvement methodology such as PRO2PI. As recent, I mean the last four years, from May 2014 to June 2018. Two of them are more related with structure based on new concepts. They are already mentioned in the introduction of this article. We can name them as Other Process Quality Characteristic and as Specific Practices for Capability Evolution. Other three of them are more related to the practical side. We can name them as SPI with Agility, Models for Innovation and SPI Education. Finally, there is one of them more related with the need of a theory. We can name it as Theory of SPI.

Other Process Quality Characteristic comes from the recent revision of ISO/IEC 15504 as the ISO/IEC 330xx family of Standards. It defines the requirements for process measurement frameworks other than capability. ISO/IEC 33003 [3] defines requirements for process measurement frameworks. Up to that, SPI was strongly related with only process capability.

SPI has been based on the underlying process management premise, “the quality of a system or product is highly influenced by the quality of the process used to develop and maintain it” [9, 10]. A second premise is based on process capability and organizational capability maturity. Process capability is “a characterization of the ability of a process to meet current or projected business goals” [12]. Organizational capability maturity is “the extent to which an organization has explicitly and consistently deployed processes that are documented, managed, measured, controlled, and continually improved” [10]. This is the definition of organizational maturity. As the model is capability maturity and the concept of capability is already assumed, I understand that the word capability is implicit in that definition. It is well recognized that those premises are based on the principles of statistical quality control by Shewhart, refined by Deming, Crosby and Juran, and applied to software by Humphrey, Radice and others.

With Process Measurement Framework, SPI is no longer necessarily related with process capability. Process capability is now a possible Process Measurement Framework. ISO/IEC 33003 mentions process capability, process security, process agility and process safety, as examples of Process Measurement Framework. ISO/IEC 33020 defines a process measurement framework for the assessment of process capability, conformant with the requirements of ISO/IEC 33003. Within this process measurement framework, the measure of capability is based upon a set of process attributes. Each process attribute defines a measurable property of process capability.

In order to consider Other Process Quality Characteristic, PRO2PI Methodology has already a starting direction: The modeling view of SPI with three types of profiles. Using this starting direction, there is a need to review two components: MMOD and MFMOD. Currently, MFMOD has only a version of a metamodel for Process Capability Profile. This metamodel needs to model the definition and utilization of different process measurement framework instead of having only the process capability metamodel, as it is now. The MFMOD needs to include orientation do define or choose a process measurement frameworks to be used in a reference model, instead of using only process capability framework, as it is now.

Specific Practices for Capability Evolution means the way two most recent versions of relevant reference models define the evolution of processes. As already mentions, they are the distinct process practices of generic profile group for Very Small Entities (VSE) from ISO/IEC 29110 [6, 7] and distinct practices for each capability level of each practice area in the recent version of CMMI (CMMI Development V2.0) [8], to characterize process evolution towards capability, instead of generic capability practices, as in previous versions of CMMI and in ISO/IEC 15504. ISO/IEC 330xx expects generic practice for a measurement framework.

In order to consider Specific Practices for Capability Evolution, PRO2PI Methodology needs to review two components: MMOD and MFMOD. Currently, MMOD model specific practices for each process and the utilization of generic practices for measurement levels [25]. MMOD needs to be changed to combine both possibilities: specific or generic practices for measurement levels. The MFMOD needs to include orientation do define or choose a combination of specific or generic practices for measurement framework for a reference model.

SPI with Agility is about integrating agility, lean and other in SPI. Agility has been around since the launching of the manifesto for agile software development in 2001. So the term agility is used here as expressed in this manifesto. The manifesto proposed four agile values (individuals and interactions, working software, customer collaboration and responding to change) over traditional correspondent ones, considered more associated with SPI (processes and tools, comprehensive documentation, contract negotiation and following a plan). The signatories of the manifesto declared that while there is value in these traditional items, they value more the agile items. At that time, agile manifesto was an alternative to SPI.

Since then, we have seen practical success and broad dissemination of agility, continuity of practical success and even broader dissemination of SPI and many practical applications of integrated agile and SPI. Then, there have been efforts to understand this integration. In a CMU/SEI technical report [33], the authors claim that although agile development methods and CMMI best practices are often perceived to be at odds with each other, there are benefits from using both. They propose that CMMI and Agile champions work together. The recent version of CMMI (CMMI Development V2.0) declares a further integration with agile

There is no need to a significant additional revision of PRO2PI Methodology in order to consider SPI with Agility. A recent article by Kuhrmann et al. [22] supports SPI with Agility. They present the results of a combined Systematic Literature Review and Systematic Mapping Study (SLR/SMS) on SPI. The objectives were to capture the domain of SPI, to provide a snapshot of the available publication pool, and to investigate research trends. 18,686 publications were identified from 1989 to 2015. They selected and analyzed 769 publications.

Among the results, they classified the 769 publications in terms of categories of research type, contribution type and main focus. They identified four research trends in SPI. These research trends and the number and percentage of articles, based on the main focus of each one, are: New or customized SPI models (295) (38%), SPI success factors (126) (16.4%), SPI for SMEs (116) (15.1%) and SPI and agility (73) (9.5%).

In a recent analysis, sixteen maturity models related to Research, Development and Innovation were identified over a period of sixteen years [2002–2017] [23]. There are two models in the first four-year period [2002–2005], three models in the period [2006–2009], four in [2010–2013] and finally seven models in the last period [2014–2017]. This significant increase of number of models in the last period, and the fact that most of them are for innovation, indicates the increasing relevance of this theme (Models for Innovation).

There is no need to significant additional revision of PRO2PI Methodology in order to consider Models for Innovation, as demonstrated with how it was used to develop CERTICS and CERTICS-SAE models.

SPI Education means the challenging effort of education of Software Process Improvement. A recent international workshop, for example, focused on the new challenges for and best practices in software process education, training and professionalism [32]. In this workshop, all articles mention this challenging effort. One of them, for example, a systematic mapping study on SPI education concludes: “in spite of its [SPI] importance, increasing its coverage in educational settings is still challenging” [32, pp. 7–17]. Another article introduces a research to understand SPI education oriented to software industry needs [32, pp. 70–74].

As described in the recent utilization of PRO2PI Methodology, the PRO2PI-WORK4E already covers SPI Education. There is no need to significant additional revision of PRO2PI Methodology in order to consider SPI Education. Of course, it does not solve SPI Education, but only provide one way to deal to teach SPI.

SPI with Agility, Models for Innovation and SPI Education may need different Process Measurement Frameworks. They should progress towards maturity following different specific process quality characteristics. These will be supported by PRO2PI Methodology with its evolution for Other Process Quality Character.

Finally, there is the challenge of providing a Theory of SPI. Kuhrmann et al. [22] supports this challenge. They concluded their SLR/SMS on SPI stating “there is a lack of discussions and critical comparisons of the approaches in practice and few on theories and models of SPI”. “Although SPI is around for decades, we still miss a sound theory about SPI” [22, p. 26].

David Card already identified the reasons for this missing sound theory of SPI in 2004: “SPI has become a driving force in the global software industry. However, it has not become a popular topic of rigorous research, especially at universities. […] [SPI] approaches have evolved or been adapted to software engineering largely without the participation of the academic research community. […] One issue that inhibits the deployment of these approaches today is that these approaches are considered competitors. In reality they are all based on very similar concepts and techniques. The packaging obscures the underlying principles. Eliciting and refining underlying principles is the role of science.” [24]

PRO2PI Methodology needs a theory of SPI. In order to pursue a theory of a modeling driven SPI, there is a need to understand the concept of theory itself. Shirley Gregor, for example, examines the structural nature of theory in Information Systems [26]. She addresses issues of causality, explanation, prediction, and generalization that underlie an understanding of theory. Then she proposes a taxonomy to classify information systems theories with respect to the manner in which four central goals are addressed: analysis, explanation, prediction, and prescription. Finally, she identifies five interrelated types of theory: (1) theory for analyzing, (2) theory for explaining, (3) theory for predicting, (4) theory for explaining and predicting, and (5) theory for design and action.

Even though I generalize SPI from software to knowledge workers processes, it still has strong connections with Software Engineering. Therefore concerns about a theory of software engineering should be considered. Johnson et al., for example, proposes an effort towards a theory of software engineering [34]. Jacobson and Meyer [35] presented steps towards a theory of software engineering and its validation: model de nature of methods (for software engineering), find the kernel (the mother of all methods) and describe each interesting method using the kernel. These steps had been used to produce the SEMAT (Software Engineering Methods and Theory) Kernel, as the essence of Software Engineering [36].

PRO2PI is a methodology for modeling driven SPI in the meaning of Model Driven Engineering (MDE). Ty, for example, explains the relationship between “model-based” and “model-driven” [27], in this case on a clearer target, a function. He suggested defining model-based in terms of “a function is model-based if it is based on the model(s) of the functional target, i.e. the thing that will be influenced by the function; thus, say a system is model-based if its major functions are model-based”. Then, he suggested defining model-driven in terms of “a model-based function/system is model-driven, if the model is changeable in the system at runtime, e.g., allows to change it when the function is executing or before each execution”. Therefore “model-based” is a special case of “model-driven”.

There are efforts to a better understanding of MDE with research towards a model theory of MDE. Favre [28], Bézivin [29], Seidwitz [30] and Muller et al. [31], for example, provide insights for a model theory of MDE. These results should be used for a model theory of modeling driven SPI.

5 Conclusion

Among the six identified recent challenges and changes in SPI context, Other Process Quality Characteristic is certainly the one that has the most impact in the practical and research context, causing the most significant expansion of SPI.

Analysis of the PRO2PI methodology, as an example of a methodology for SPI, in relation to these six identified recent challenges and changes in SPI context, especially Other Process Quality Characteristic, indicates two directions. First, it corroborates PRO2PI as a promising methodology. Even without being developed with the Other Process Quality Characteristic prediction, PRO2PI already has a concept to be evolved based on the multiple types of profiles from the SPI modeling view. For Specific Practices for Capability Evolution, and Other Process Quality Characteristic there is a need to update basically two PRO2PI components: MMOD and MFMOD. The most recent utilization and the fact of already supporting some of the identified challenges and changes, in this case, Models for Innovation and SPI Education, also helps in this corroboration. For SPI with Agility, together with Models for Innovation and SPI Education, there is no need to update PRO2PI. The need for a Theory of SPI was already identified for PRO2PI.

Second direction is that, even though PRO2PI already has a basis, it needs to be revised to incorporate evolutions to fully support three of the six challenges and changes: Other Process Quality Characteristic, Specific Practices for Capability Evolution and Theory of SPI. The fact that Other Process Quality Characteristic has been defined in ISO/IEC 33003 based on a solid theory of measurement, backed by the practice of SPI, indicates that there is sufficient practice for a theory of SPI. The search for model theory for MDE and theory for software engineering indicate a path to a model theory of modeling driven SPI.

A limitation of this research is that the identification of these six challenges and changes did not follow a systematic approach. They are identified using subjective experience and observation. They are also influenced by the needs and characteristics of PRO2PI Methodology, as an example of SPI Methodology. Each one of them, however, is supported by independent objective results. Other Process Quality Characteristic is introduced by ISO/IEC 33003 [3], a major player in SPI context. Two major reference models introduce specific Practices for Capability Evolution: CMMI V2.0 [8] and ISO/IEC 29110 [6]. Models for Innovation is supported by the increase in the number of reference models for that [23]. SPI Education is supported by a workshop on that subject [32]. SPI with Agility, and a Theory of SPI are supported by a recent comprehensive SLR/SMS on SPI [22].

This article is more a critical reflection on relevance recent challenges and changes in SPI research and practical context. The presented impact of these recent challenges and changes in PRO2PI Methodology is an example. Each SPI methodology should be reviewed and evolved considering these six and possibly other challenges and changes in the context of SPI to provide better support for SPI.