Together with many success stories, promises such as the increase in production speed and the improvement in stakeholders’ collaboration have contributed to making agile a transformation in the software industry in which many companies want to take part. However, driven either by a natural and expected evolution or by contextual factors that challenge the adoption of agile methods as prescribed by their creator(s), software processes in practice mutate into hybrids over time. Are these still agile? In this article, we investigate the question: what makes a software development method agile? We present an empirical study grounded in a large-scale international survey that aims to identify software development methods and practices that improve or tame agility. Based on 556 data points, we analyze the perceived degree of agility in the implementation of standard project disciplines and its relation to used development methods and practices. Our findings suggest that only a small number of participants operate their projects in a purely traditional or agile manner (under 15 percent). That said, most project disciplines and most practices show a clear trend towards increasing degrees of agility. Compared to the methods used to develop software, the selection of practices has a stronger effect on the degree of agility of a given discipline. Finally, there are no methods or practices that explicitly guarantee or prevent agility. We conclude that agility cannot be defined solely at the process level. Additional factors need to be taken into account when trying to implement or improve agility in a software company. Finally, we discuss the field of software process-related research in the light of our findings and present a roadmap for future research.

Driven by the need to coordinate activities of multiple agile development teams cooperating to produce a large software product, software-intensive organizations are turning to scaling agile software development frameworks. Despite the growing adoption of various scaling agile frameworks, there is little empirical evidence of how effective their practices are in mitigating risk, especially in global software development (GSD), where project failure is a known problem. In this study, we develop a GSD Risk Catalog of 63 risks to assess the degree to which two scaling agile frameworks–Disciplined Agile Delivery (DAD) and the Scaled Agile Framework (SAFe)–address software project risks in GSD. We examined data from two longitudinal case studies implementing each framework to identify the extent to which the framework practices address GSD risks. Scaling agile frameworks appear to help companies eliminate or mitigate many traditional risks in GSD, especially relating to users and customers. However, several important risks were not eliminated or mitigated. These persistent risks in the main belonged to the Environment quadrant highlighting the inherent risk in developing software across geographic boundaries. Perhaps these frameworks (and arguably any framework), would have difficulty alleviating, issues that appear to be outside the immediate control of the organization.

Background: Agile methods are no longer restricted to small projects and co-located teams. The last decade has seen the spread of agile into large scale, distributed and regulated domains. Many case studies show successful agile adoption in GSD, however, taken as a whole, it remains unclear how widespread this trend is, and what form the agile adoption takes in a global software development (GSD) setting. Aims: Our objective is to gain a deeper understanding of how organisations adopt agile development methods in distributed settings. Specifically we aim to plot the current development process landscape in GSD. Method: We analyse industrial survey data from 33 different countries collected as part of the HELENA project that explored the wider use of hybrid development approaches in software development. We extract and analyse the results of 263 surveys completed by participants involved in globally distributed projects. Results: In our sample, 72% of globally distributed projects implement a mix of both agile and traditional approaches (termed `hybrid'). 25% of GSD projects are predominantly agile, with only very few (2%) opting for traditional approaches. GSD projects that used only agile methods tended to be very large. Conclusions: Globally Distributed Software Development (and project size) is not a barrier to adopting agile practices. Yet, to facilitate project coordination and general project management, many adopt traditional approaches, resulting in a hybrid approach that follows defined rules.

It is becoming commonplace for companies to contribute to open source software (OSS) projects. At the same time, many software organizations are applying Scrum software development practices, for productivity and quality gains. Scrum calls for self-organizing teams, in which the development team has total control over its development process. However, OSS projects typically have their own processes and standards, which might not mesh well with a company's internal processes, such as Scrum. This paper presents an experience report from Sony Interactive Entertainment (SIE), where the "toolchain CPU compiler" team directly participates in the "LLVM" OSS project. The team ran into a number of difficulties when using Scrum to manage their development. In particular, the team often failed to complete Scrum sprints where tasks required interaction with the open source community. We look at how the team redefined task flows to alleviate these difficulties, and eventually evolved a highly modified process, dubbed TWINS (This Workflow Is Not Scrum). We assess the revised process, and compare it to other established agile methods, finding it bears a strong resemblance to Scrumban (the SIE team was not aware of Scrumban previously). The TWINS framework presented here may help other organizations who develop software in-house and engage in OSS projects, to gain the best of both worlds.

We are pleased to introduce this Special Issue on Global Software Engineering Education published by the ACM Transactions on Computing Education (TOCE) that focuses on educational practices to prepare students for a global workplace. This issue comes at a time when universities are recognizing the need to provide courses that address the challenges of distributed development and presents research that will facilitate course leaders currently running, or embarking on, Global Software Engineering Education (GSE-Ed).

High staff turnover has a negative impact on software development productivity and product quality. Further, offshore outsourcing has a widely held reputation for particularly poor employee retention. Interestingly, in-house sites (regardless of location) do not suffer such high levels of staff turnover.

The adoption, scaling and tailoring of agile methods depends on several factors, such as the size of the software development organization, business goals, and operative model. The Scaled Agile Framework (SAFe) was developed to support organizations in scaling agile practices across the enterprise.

Distributed development involving globally distributed teams in different countries and timezones adds additional complexity into an already complex undertaking. This paper focuses on the effect of global software development on motivation. Specifically, we ask, what impact does misalignment between needed and actual autonomy have on global team motivation?

Global Software Engineering (GSE) is a reality for even the smallest companies, so software engineering students need to learn how to work in a globally distributed development context. Many approaches to teaching GSE have been described in the literature. Since the majority of software development is done by engineers working in small or medium sized enterprises (SMEs) we now ask: Are today's students being trained to work effectively in small distributed companies? We surveyed three GSE SMEs to identify which of 70 Global Teaming Model (GTM) practices were problematic and important to this sample. We then mapped recommendations for GSE educators to those pinpointed GTM practices. Finally, we analysed the level to which these needed GTM practices were addressed by the GSE-Education (GSE-Ed) literature, and who performed these practices. Nine GTM practices were found important and relevant to all three SMEs. Seven of these were addressed by GSE-Ed recommendations, and two were seen to be lacking. A rich set of 63 unique GSE-Ed recommendations were found to support the seven GTM practices, but our analysis unearthed a surprising complexity of roles and responsibilities undertaken by the instructor in GSE-Ed courses. As a result student and client involvement in coordination and collaboration activities tended to be weakened or non-existent. In order to ensure graduates are prepared for the reality, practitioners of SMEs need to take on a more active role in the education process. Also, students need to be given more responsibility so they can learn the broader professional and management skills required when developing software in multi-site SME teams.

An ITiCSE 2015 conference working group report has recently delivered a comprehensive systematic review of the literature addressing the challenges and solutions in teaching global software engineering courses. In this panel session, a group of authors of that report will review the findings, and present their own views on teaching such courses, the pros and cons and their own motivations and experiences, together with strategies that have proven effective. We argue that these courses are critical for preparing students for practice as software engineers in today's global settings, but they are certainly not for the faint-hearted. After the review and presentation of position statements by panelists there will be a question and answer session with the audience and a discussion elaborating on the next frontiers in research and practice for global software engineering education.

Global software development (GSD) is increasingly the norm. As firms expand into new markets overseas, acquire companies in different countries, and recruit talented developers in new locations, projects become distributed by necessity. As projects become increasingly distributed, and include external participants such as outsourcing vendors, conventional top-down management oversight and control becomes more difficult. How can organizations ensure that the activities of all parties involved are aligned with the strategic objectives and values of a software development undertaking? The Global Teaming Model is a framework that specifies practices and recommendations for Global Software Development. As such, it serves as a means to organize GSD practices relevant to Software Development Governance. We conducted a case study of a small team engaged in Global Software Development, identifying governance shortcomings. Then, we used the Global Teaming Model to identify governance practices that would address those shortcomings. We identified several aspects of the team's interactions with other teams in the company that would be improved by recommendations from the Global Teaming Governance Model. The Global Teaming Model provides a blueprint for Software Development Governance that organizations seeking to globalize their current development projects in a controlled way can use to implement good Software Development Governance.

In this paper we propose a new framework to assess survivability of software projects accounting for media capability details as introduced in Media Synchronicity Theory (MST). Specifically, we add to our global engineering framework the assessment of the impact of inadequate conveyance and convergence available in the communication infrastructure selected to be used by the project, on the system ability to recover from project disasters. We propose an analytical model to assess how the project recovers from project disasters related to process and communication failures. Our model is based on media synchronicity theory to account for how information exchange impacts recovery. Then, using the proposed model we evaluate how different interventions impact communication effectiveness. Finally, we parameterize and instantiate the proposed survivability model based on a data gathering campaign comprising thirty surveys collected from senior global software development experts at ICGSE'2014 and GSD'2015.

After more than a decade of research in Global Software Engineering (GSE), organisations have a wealth of practices that they can draw on to support them in their global development activities. However, practitioners are now asking, "What is the current status of my GSE practices?", "How prepared is my company for GSE?", "What practices need improving?" We aim to give practitioners the answers they need by giving them a snapshot of their current GSE process strengths and weaknesses that can act as a guide for future software process improvement activities. We do this through our lightweight Global Teaming Assessment (GTA) based on 70 recommended GSE practices which we piloted in a small company.

Abstract The quality of an organization's support services can be key to maintaining and extending its customer base and to its future survival. As such, support services need to be considered in software process improvement initiatives and not left as an afterthought or stand‐alone effort. However, to identify areas in need of improvement requires a good understanding of the current process. To meet this need, we have developed a process modeling technique called LAPPI that documents current ‘as‐is’ processes. This study explores whether the LAPPI can identify opportunities for improvement in the customer service process. LAPPI is applied, step by step, in a small‐to‐medium‐sized enterprise. Results are validated through an analysis of customer satisfaction questionnaires administered before and after our LAPPI intervention. Applying the LAPPI highlighted several problems with the customer support service, which was subsequently streamlined. The validation shows an improvement in customer satisfaction in several areas, with significant improvement in customer response times and in customer query resolution. Software development organizations can use LAPPI to highlight where improvements are needed in support services. This study shows that making best use of support resources and ensuring customers receive prompt and clear help throughout the product's lifetime can improve an organization's image and future prosperity. Copyright © 2015 John Wiley & Sons, Ltd.

Background. In order to understand research on a particular computing topic, practitioners and researchers often need to obtain an overview of its research methods. Current research methods coding schemes either capture insufficient details to support a full critical assessment, or are specialised to a particular research type. Aim. This paper defines and demonstrates RSML, a Research Schema Modelling Language that captures a high level of detail and is applicable to most types of computing research. Method. RSML was designed using concepts from the research methods literature, then refined inductively. An RSML editor was created to assist coders and help reduce coding errors. To demonstrate the feasibility of modelling research with RSML, and to exemplify the summary information that can be derived from a database of RSML encodings, a trial review of 24 articles from one journal was conducted. Results. The review illustrates quantitatively the journal's focus on artifact construction and empiricism. It also reveals that observations are rarely used to inform artifact construction, and purely empirical studies are scarce. Conclusion. RSML can be used to model sophisticated, multifaceted research spanning a wide range of software engineering topics, yielding insights that are not easily captured by current coding schemes.

Context: One of the chief challenges of Global Software Development (GSD) is for globally, culturally and linguistically diverse team members to communicate effectively. Failing to meet this challenge can lead to misunderstandings that impede project success. To prepare practitioners for remote and diverse forms of communication we have developed VENTURE: a simulation-based training platform that aims to increase cultural awareness.

The evaluation of educational approaches is frequently challenging as it involves humans and requires the consideration of several angles depending on the educational objectives. On occasions, these objectives involve academic and industrial environments, thus implying additional challenges. Education in Global Software Development (GSD) has recently became important for companies that offer GSD training in specific GSD areas such as cultural, linguistic, leadership, negotiation or communication skills. We have developed a training framework based on simulation with the aim of providing GSD interaction training. This framework addresses various kinds of problems related to communication barriers. However, discovering the needs and requirements of both academia and industry is a key success factor when seeking an effective educational solution. This paper presents a multi-phased methodology that has been employed to: 1) research the training requirements in the field; 2) design the training framework based on these findings; 3) evaluate the adequacy of the framework using an Expert Feedback study and a subsequent Heuristic Evaluation; 4) prove that the framework can actually generate learning; 5) conduct a market-focused study to evaluate the commercial options of the framework and 6) define future work to fulfil commercial requirements and improve effectiveness.

Software projects that are engineered using global software development techniques are required to deploy processes and tools to support collaboration over large geographies. Specifically, these projects involve the deployment of processes and tools to support project management, communication, and risk management. Whereas a traditional software development effort can use standard processes and tools to support communication and collaboration, global software development efforts require a unified and holistic project management, development process, collaboration, and communication approach taking into consideration the interplay of time zones, number of sites and cultural diversity. In this paper, we introduce a novel approach for modeling and quantification of global software engineering frameworks. In our approach, we apply transient survivability metrics to support the design of global software engineering projects. Therefore, our approach combines survivability analysis and global software engineering frameworks analysis. The survivability metric used in this paper is the time required to recover from a software project disaster (e.g., break down in communication between remote teams) for a given global software engineering framework. The global software engineering modeling framework we use is composed of models supporting the evaluation of communication tools, software development processes and cultural diversity management. We illustrate the application of our approach by applying it to the analysis of an example derived from a real global software engineering project. Our results indicate that the combination of survivability analysis and modeling of global software engineering frameworks can provide meaningful insights when designing global software engineering frameworks.

Universities are starting to offer courses in Global Software Development (GSD) as a way to prepare future practitioners for challenges found in organizational settings. Companies also provide GSD training especially to new recruits who might lack skills in how to effectively communicate with people from different cultures and different first languages. However this training is costly and not necessarily tailored to the specific needs of the organization or individual practitioner. With a focus on GSD education, we have developed a training framework based on simulation that addresses different kinds of GSD problems. The framework allows course leaders to design training scenarios to fit their precise needs for improved communication and cultural awareness. This paper presents a market-focused study aimed at discovering the requirements for a commercial GSD education tool. We studied the market options, identified key stakeholders, and conducted interviews with senior managers from three multinational companies. The study helped us to gain a deeper understanding of current training trends, needs and gaps according to key stakeholders. Finally, outcomes provided us with directions for future tool development. Results indicate that industry find cultural differences a more pressing problem than linguistic differences. Although industry tends to use traditional (classroom based) training methods for raising cultural awareness, they perceive that interaction simulation is a viable alternative. Performance and potential cost savings are key drivers for commercialising prototype tools. Conducting market-focussed research can help to ensure our future research is relevant, and meets the needs of industry.

Context: More than a decade of empirical research into global software development has yielded a wealth of solutions to recurring problems. The Global Teaming Model is one example, organizing extensive empirical research into a cohesive process model for managing global software development projects. Problem: evidence shows that empirical research does not reach practitioners: software project managers do not read academic publications. Thus research results need to be cast in a form that is accessible to practitioners. Approach: We employ a two-pronged approach to address this problem. First, we translate research results into software process patterns, that describe a specific software development problem and an empirically derived and validated solution. Second, we enter the patterns into a knowledge based system that recommends patterns applicable to a software development project. Result: We used this approach to create a decision support system based on the Global Teaming Model. The result makes the sixty specific practices in the Global Teaming Model immediately accessible to managers and developers involved in Global Software Development. Also, the patternization of Global Teaming Model practices proved to be an efficient, low-overhead approach to translating research results into a practitioner-friendly format. Conclusion: The combination of patterns and knowledge-based decision support promises to bridge the research-practice divide by making research results more accessible to those who might benefit from them.

When it comes to software process improvement (SPI), process engineers look for SPI methods to support process analysis, design, realization, deployment, and management. Although a number of different SPI methods and models exist, process engineers tend to view these as too generic, too large, or a poor fit for the organization in which SPI is conducted. A strategy to overcome these shortcomings is to concentrate on the artifacts, which precisely define the desired outcomes, rather than on specific methods. In this paper, we present the Artifact-based Software Process Improvement & Management (ArSPI) model that provides a unified perspective on SPI and company-wide software process management (SPM), the required key artifacts, and the life cycle models. ArSPI is shown to be of practical support to industry who called for a practical way to define the interfaces between SPI projects. This paper concludes with an example of how ArSPI paved the way for several organizations through applying the model in real-world SPI-projects.

Simulation has been applied in several Software Engineering fields, and is shown to be a useful method in industrial training. As part of our research work, we have used simulation to provide training in Global Software Development (GSD). We have developed a platform to strengthen GSD skills by simulating realistic settings in which learners interact with Virtual Agents of differing cultures. Thus, learners will experience multi-cultural problems and will develop specific GSD communication skills. The development of these skills must, however, be accurately assessed, bearing in mind that the training is aimed at learners with different characteristics and skills. In this paper we present an assessment process based on educational theory adapted to our simulation-based training environment. Methods to minimize the instructors' effort and tailor the assessment to specific training needs are proposed. The assessment process has been evaluated by 34 potential users. Results indicate that the assessment method yields meaningful results and proof of learning. Also that automated assessment can be achieved with minimal intervention from the instructor. Though tailored for GSD, this method could be applied to other domains, in both industry and academia.

Content analysis is a useful approach for analyzing unstructured software project data, but it is labor-intensive and slow. Can automated text classification (using supervised machine learning) be used to reduce the labor or improve the speed of content analysis? We conducted a case study involving data from a previous study that employed content analysis of an open source software project. We used a human-coded data set with 3256 samples to create different size training sets ranging in size from 100 to 3000 samples to train an "ensemble" text classifier to assign one of five different categories to a test set of samples. The results show that the automated classifier could be trained to recognize categories, but much less accurately than the human classifiers. In particular, both precision and recall for low-frequency categories was very low (less than 20%). Nevertheless, we hypothesize that automated classifiers could be used to filter a sample to identify common categories before human researchers examine the remainder for more difficult categories.

Twelve years ago a group of practitioners and researchers came together to try to solve problems relating specifically to Global Software Engineering (GSE) practice. This paper aims to assess whether the many hundreds of GSE research papers written over this period have had an impact on practice. We conducted semi-structured interviews with senior managers and project managers from ten companies, four of which are large multinationals (three in Fortune 100), four are medium sized enterprises, and two are small startups. GSE research is perceived as useful by industry with all participants stating that studying the subject would improve GSE performance, but all were unanimous in saying they did not read articles on GSE. Practitioners go to books, blogs, colleagues, forums, experience reports of 1-2 pages in length, or depend on their own experience to solve problems in GSE. Controversially, many didn't see GSE as separate from general project management. Practitioners don't want frameworks, they want patterns of context specific help. While dissemination techniques need to be improved, that is not sufficient. Experience-based advice is just as important.

Global software development (GSD) carried out at various locations is impeded by global distance categorized as exterior and interior. To overcome this, specific communication strategies are required to coordinate a range of activities between client and vendor teams working on GSD projects. The literature on GSD recommends many coordination strategies, but tends to take the client viewpoint as vendor companies have been underrepresented in publications. In this paper, we compare coordination strategies identified in both the literature and in primary research which we conducted in Indian vendor companies with the Project Communication Management section of the PMBOK® Guide. This comparison helped us to define a set of expert-validated coordination strategies for use by GSD practitioners. While developed from the vendor rather than the client perspective, indications are that project managers from both client and vendor organisations can implement these strategies to overcome GSD coordination issues.

ABSTRACT Understanding the current state of the software processes and their problem points is important. Without this understanding, software process improvement (SPI) resources may be allocated to less meaningful targets. SPI work can be challenging to initiate especially in small companies where resources and knowledge of SPI are often limited. The aim of the developed technique, LAPPI (A Light‐weight Technique to Practical Process Modeling and Improvement Target Identification), is to provide an easy to use, lightweight tool for process modeling and improvement target identification. The technique provides a suitable method that integrates with various SPI initiatives. The method used in the development of LAPPI is a nonformal variation of constructive research. LAPPI has been incrementally developed in multiple academia‐industry collaboration projects and by industry actors themselves. Our evaluation of the LAPPI technique in 42 studies conducted in 31 companies indicates that the technique is suitable for modeling the current process and identifying the points of improvement in the process. Practical experience shows that LAPPI provides a cost‐effective technique for process modeling and improvement target identification especially in small and medium‐sized enterprises. It is most useful in the diagnosing phase of SPI. It helps the company to understand the current processes and the organizational interactions, and to create a process description baseline. Copyright © 2012 John Wiley & Sons, Ltd.

Background: The accurate prediction of where faults are likely to occur in code can help direct test effort, reduce costs, and improve the quality of software. Objective: We investigate how the context of models, the independent variables used, and the modeling techniques applied influence the performance of fault prediction models. Method: We used a systematic literature review to identify 208 fault prediction studies published from January 2000 to December 2010. We synthesize the quantitative and qualitative results of 36 studies which report sufficient contextual and methodological information according to the criteria we develop and apply. Results: The models that perform well tend to be based on simple modeling techniques such as Naive Bayes or Logistic Regression. Combinations of independent variables have been used by models that perform well. Feature selection has been applied to these combinations when models are performing particularly well. Conclusion: The methodology used to build models seems to be influential to predictive performance. Although there are a set of fault prediction studies in which confidence is possible, more studies are needed that use a reliable methodology and which report their context, methodology, and performance comprehensively.

Open Source software is competing successfully in many areas. The commercial sector is recognizing the benefits offered by Open Source development methods that lead to high quality software. Can these benefits be realized in specialized domains where expertise is rare? This study examined discussion forums of an Open Source project in a particular specialized application domain - electronic medical records - to see how development roles are carried out, and by whom. We found through a qualitative analysis that the core developers in this system include doctors and clinicians who also use the product. We also found that the size of the community associated with the project is an order of magnitude smaller than predicted, yet still maintains a high degree of responsiveness to issues raised by users. The implication is that a few experts and a small core of dedicated programmers can achieve success using an Open Source approach in a specialized domain.

Architectural Knowledge Management (AKM) aims to coordinate the knowledge produced and used during architecting a software system. Managing architectural knowledge effectively is a task that becomes even more critical and complex when operating in a distributed environment. Thus, software architectural practices, processes, and tools that work in collocated software development don't necessarily scale up in a distributed environment. In this paper, we perform a literature review that looks at AKM in a Global Software Development (GSD) context. We attempt to synthesize AKM concepts, practices, tools and challenges important in GSD. In order to provide a common understanding for the central concepts of AKM in GSD in an abstract way, we have created a metamodel which is based on our literature review. The metamodel defines a set of architecture knowledge and global software development entities and their relationships.

In this paper, we present the Global Teaming Model (GTM), which is empirically grounded, and outlines practices that managers need to consider when managing virtual teams. We explain how the model can be adapted to specific areas of software development, and use architectural knowledge management (AKM) as our exemplar. We focus on specific practices relating to how teams collaborate and share essential architectural knowledge across multiple sites. Through a review of the literature, we develop an in-depth view of recommended practices associated with AKM in a global environment. We then consider how we can incorporate these AKM practices into our Global Teaming model to ensure managers are given the necessary support. Our contribution to research therefore is to present AKM practices within the context of all other Global Software Development processes.

Cultural diversity is assumed to be a fundamental issue in global software development. Research carried out to date has raised concerns over how to manage cultural differences in global software development. Our empirical research in India, a major outsourcing destination, has helped us investigate this complex issue of global software development. A triangulated study based on a questionnaire, telephonic interviews and structured face-to-face interviews with 15 Project Managers and Senior Executives has revealed how they cope with the demands of cultural differences imposed by a geographically distributed environment. This research study brings forward various techniques initiated by these project managers to deal with cultural differences that exist within geographically distributed software development teams. We also discuss different strategies and make a case to explain how to build on and take advantage of cultural differences that exist in global software development.

Motivated software engineers make a critical contribution to delivering successful software systems. Understanding the motivations of software engineers and the impact of motivation on software engineering outcomes could significantly affect the industry's ability to deliver good quality software systems. Understanding the motivations of people generally in relation to their work is underpinned by eight classic motivation theories from the social sciences. We would expect these classic motivation theories to play an important role in developing a rigorous understanding of the specific motivations of software engineers. In this article we investigate how this theoretical basis has been exploited in previous studies of software engineering. We analyzed 92 studies of motivation in software engineering that were published in the literature between 1980 and 2006. Our main findings are that many studies of software engineers' motivations are not explicitly underpinned by reference to the classic motivation theories. Furthermore, the findings presented in these studies are often not explicitly interpreted in terms of those theories, despite the fact that in many cases there is a relationship between those findings and the theories. Our conclusion is that although there has been a great deal of previous work looking at motivation in software engineering, the lack of reference to classic theories of motivation means that the current body of work in the area is weakened and our understanding of motivation in software engineering is not as rigorous as it may at first appear. This weakness in the current state of knowledge highlights important areas for future researchers to contribute towards developing a rigorous and usable body of knowledge in motivating software engineers.

In this paper, we describe our investigation of the motivational differences between project managers and developers. Motivation has been found to be a central factor in successful software projects. However the motivation of software engineers is generally poorly understood and previous work done in the area is thought to be largely out-of-date. We present data collected from 6 software developers and 4 project managers at a workshop we organized at the XP2006 international conference.

Anecdotal evidence from our work with software developers suggests that maintenance is a significant problem for software development companies. A problem that is absorbing increasing amounts of precious development effort. In parallel, software companies are increasingly applying process improvement principles to development problems. In this paper we discuss how maintenance is addressed in process improvement programmes. We look at how well maintenance is addressed by formal process models like CMM. We also present empirical evidence from our study of process improvement in UK software companies. Our main findings are that although developers report that maintenance is indeed a problem, it is not always their most important problem. Furthermore, our findings also suggest that companies are often not well prepared for the maintenance phase of developments and that formal process improvement models do not pay enough attention to maintenance.