Product-Focused Software Process Improvement

As happened previously in electronics, software targeted at consumers is increasingly the focus of investment and innovation. Some of the areas where it is leading is animated interfaces, treating users as a community, audio and video information, software as a service, agile software development, and the integration of business models with software design. As a risk-taking and experimental market, and as a source of ideas, consumer software can benefit other areas of applications software. The influence of consumer software can be magnified by research into the internal organizations and processes of the innovative firms at its foundation.

Software “best” practices depend entirely on context – in terms of the problem domain, the system constructed, the software designers, and the “customers” ultimately deriving value from the system. Agile practices no longer have the luxury of “choosing” small non-mission critical projects with co-located teams. Project stakeholders are selecting and adapting practices based on a combina tion of interest, need and staffing. For example, growing product portfolios through a merger or the acquisition of a company exposes legacy systems to new staff, new software integration challenges, and new ideas. Innovation in communications (tools and processes) to span the growth and contraction of both information and organizations, while managing the adoption of changing software practices, is imperative for success. Traditional web-based tools such as web pages, document libraries, and forums are not suf ficient. A blend of tweeting, blogs, wikis, instant messaging, web-based confer encing, and telepresence creates a new dimension of communication “best” practices.

The information about which modules of a future version of a software system are defect-prone is a valuable planning aid for quality managers and testers. Defect prediction promises to indicate these defect-prone modules. However, constructing effective defect prediction models in an industrial setting involves a number of key questions. In this paper we discuss ten key questions identified in context of establishing defect prediction in a large software development project. Seven consecutive versions of the software system have been used to construct and validate defect prediction models for system test planning. Furthermore, the paper presents initial empirical results from the studied project and, by this means, contributes answers to the identified questions.

Different Software Requirements Specifications (SRS) are hard to compare due to the uniqueness of the projects they were created in. Without such comparison, it is difficult to objectively determine if a project’s SRS is good enough to serve as a foundation for project success. We define a quality model for SRS and derive required metrics using the Goal-Question-Metric approach. These metrics were applied in roughly 40 student’s software projects. Based on this we find a quality threshold for project success. This paper contributes in three areas: Firstly, we present our quality model. It was derived from literature, and contributes to the discussion of how to objectively measure requirements quality. Secondly, we share our evaluation approach and our experiences measuring SRS quality. Others could profit, when planning to measure requirements quality. Finally, we present our findings and compare them to related studies in literature.

There has been number of measurement techniques proposed in the literature. These metrics can be used in assessing quality of software products, thereby controlling costs and schedules. The empirical validation of object-oriented (OO) metrics is essential to ensure their practical relevance in industrial settings. In this paper, we empirically validate OO metrics given by Chidamber and Kemerer for their ability to predict software quality in terms of fault proneness. In order to analyze these metrics we use gene expression programming (GEP). Here, we explore the ability of OO metrics using defect data for open source software. Further, we develop a software quality metric and suggest ways in which software professional may use this metric for process improvement. We conclude that GEP can be used in detecting fault prone classes. We also conclude that the proposed metric may be effectively used by software managers tin predicting faulty classes in earlier phases of software development.

There are many methods of software cost estimating (COCOMO, function points analysis, three-point estimate, use case points, class points, XP user stories, SLOC prediction and others), with their advantages and drawbacks. One common problem with all methods is the necessity to estimate the whole requirements specification, item by item. At the end, either this process is expensive or the numbers are inaccurate. This paper presents a method of software cost estimating using a limited number of functional requirements, called Scope Champions. Estimators produce more detailed and grounded numbers that are used in a final estimation formula. The method reduces the costs of estimating and increases accuracy.

When developing software-intensive products for a market-place it is important for a development organisation to create innovative features for coming releases in order to achieve advantage over competitors. This paper focuses on assessment of innovation capability at team level in relation to the requirements engineering that is taking place before the actual product development projects are decided, when new business models, technology opportunities and intellectual property rights are created and investigated through e.g. prototyping and concept development. The result is a measurement framework focusing on four areas: innovation elicitation, selection, impact and ways-of-working. For each area, candidate measurements were derived from interviews to be used as inspiration in the development of a tailored measurement program. The framework is based on interviews with participants of a software team with specific innovation responsibilities and validated through cross-case analysis and feedback from practitioners.

Reliable effort and cost estimation remains to be a challenging issue even for mature software organizations. Although, these organizations collect historical data to base their future estimates, changes in circumstances (such as application type, development platform, etc.) prevent their successful utilization. As a result, companies often suffer from underestimated and unrealistic schedules. Managing software projects that involve a large number of globally distributed stakeholders makes estimation and planning even more challenging. Related studies show that even knowledgeable project managers often underestimate hidden costs and sources of delay associated with distributed development. Therefore, management activities such as estimation of development effort, planning and control require special attention. In this paper we discuss experiences gained from a highly distributed software project, which aimed at development of a product based on a new platform and architectural solution. The project was conducted in a CMMI Level 5 company and still failed to meet initial plan constraints. We thus provide an overview of management decisions in the light of their consequences, and discuss potential areas of improvement.

This paper presents results of a literature analysis on Empirical Research Approaches in Software Engineering (SE). The analysis explores reasons why traditional methods, such as statistical hypothesis testing and experiment replication are weakly utilized in the field of SE. It appears that basic assumptions and preconditions of the traditional methods are contradicting the actual situation in the SE. Furthermore, we have identified main issues that should be considered by the researcher when selecting the research approach. In virtue of reasons for weak utilization of traditional methods we propose stronger use of Multi-Method approach with Pragmatism as the philosophical standpoint.

Technology transfer and innovation diffusion are key success factors for an enterprise. The shift to a new software technology involves, on one hand, inevitable changes to ingrained and familiar processes and, on the other, requires training, changes in practices and commitment on behalf of technical staff and management. Nevertheless, industry is often reluctant to innovation due to the changes it determines. The process of innovation diffusion is easier if the new technology is supported by empirical evidence. In this sense our conjecture is that Empirical Software Engineering (ESE) serves as means for validating and transferring a new technology within production processes. In this paper, the authors report their experience of a method, Multiview Framework, defined in the SERLAB research laboratory as support for designing and managing a goal oriented measurement program that has been validated through various empirical studies before being transferred to an Italian SME. Our discussion points out the important role of empirical evidence for obtaining management commitment and buy-in on behalf of technical staff, and for making technological transfer possible.

As agile methods and Global Software Development (GSD) are become increasingly popular, GSD project managers have been exploring the viability of using agile approaches in their development environments. Despite the expected benefits of using an agile approach with a GSD project, the overall combining mechanisms of the two approaches are not clearly understood. To address this challenge, we propose a conceptual framework, based on the research literature. This framework is expected to aid a project manager in deciding what agile strategies are effective for a particular GSD project, taking into account project context. We use an industry-based case study to explore the components of our conceptual framework. Our case study is planned and conducted according to specific published case study guidelines. We identify the agile practices and agile supporting practices used by a GSD project manager in our case study and conclude with future research directions.

Business value is a key concept in agile software development approaches. This paper presents results of a systematic review of literature on how business value is created by agile projects. We found that with very few exceptions, most published studies take the concept of business value for granted and do not state what it means in general as well as in the specific study context. We could find no study which clearly indicates how exactly individual agile practices or groups of those create value and keep accumulating it over time. The key implication for research is that we have an incentive to pursue the study of value creation in agile project by deploying empirical research methods.

Today’s software business development projects often lay claim to low-risk value to the customers in order to be financed. Emerging agile processes offer shorter investment periods, faster time-to-market and better customer satisfaction. To date, however, in agile environments there is no sound methodological schedule support contrary to the traditional plan-based approaches. To address this situation, we present an agile iteration scheduling method whose usefulness is evaluated with post-mortem simulation. It demonstrates that the method can significantly improve load balancing of resources (cca. 5×), produce higher quality and lower-risk feasible schedule, and provide more informed and established decisions by optimized schedule production. Finally, the paper analyzes benefits and issues from the use of this method.

Agile methods have appeared as an attractive alternative to conventional methodologies. These methods try to reduce the time to market and, indirectly, the cost of the product through flexible development and deep customer involvement. The processes related to requirements have been extensively studied in literature, in most cases in the frame of conventional methods. However, conclusions of conventional methodologies could not be necessarily valid for Agile; in some issues, conventional and Agile processes are radically different. As recent surveys report, inadequate project requirements is one of the most conflictive issues in agile approaches and better understanding about this is needed. This paper describes some findings concerning requirements activities in a project developed under an agile methodology. The project intended to evolve an existing product and, therefore, some background information was available. The major difficulties encountered were related to non-functional needs and management of requirements dependencies.

Refactoring, which is an efficient method to improve the quality of the existing code, has been widely used in practical software development and maintenance activities. The current refactoring researches are more focus on the technical aspect of refactoring but pay little attention to its use in real software development environment. However, software development and maintenance, in their nature, are human-centric activities. The lack of systematic empirical studies has resulted in the gap between current refactoring researches and industrial practices. To bridge this research gap, we conduct this exploratory study to learn more about the actual use of refactoring in the industrial context. Using a series of semi-controlled interviews as our major research method, we gathered first-hand information on how the refactoring is used by practitioners. We built a three-stage framework to describe the overall refactoring process. 19 basic factors are identified and categorized. We also identify the most important ones and the factors that may trigger potential conflicts between developers and the managers. Some related issues such as this study’s implications are also discussed.

In this paper we study, how to absorb software testing into the Scrum method. We conducted the research as an action research during the years 2007-2008 with three iterations. The result showed that testing can and even should be absorbed to the Scrum method. The testing team was merged into the Scrum teams. The teams can now deliver better working software in a shorter time, because testing keeps track of the progress of the development. Also the team spirit is higher, because the Scrum team members are committed to the same goal. The biggest change from test manager’s point of view was the organized Product Owner Team. Test manager don’t have testing team anymore, and in the future all the testing tasks have to be assigned through the Product Backlog.

Explaining how organizations chance has been a central and enduring quest of management scholars and many other disciplines. In order to be successful change requires not only a new process or technology but also the engagement and participation of the people involved. In this vein the change process results in new behavior and is routinized in practical daily business life of the company. Change management provides a framework for managing the human side of these changes. In this article we present a literature review on the change management in the context of Software Process Improvement. The traditional view of learning, as a “lessons learned” or post-mortem reporting activity is often apparent in SPI literature. However, learning can also be viewed as a continuous change process where specific learning cycle starts with creative conflict and ends up in formal norms and systems. Since this perspective has almost no visibility in SPI literature of past it could show a new direction to the future development of change management in SPI.

There is a need to better understand improvement of software inspection practices. Typically process improvement is driven by different reference models like CMMI. In the previous work we found that such model based approach alone didn’t provide sufficient support for inspection process improvement. This paper outlines an overall picture of inspection process improvement and four different approaches for it. These approaches are seen as complementary and the focus of the paper is on describing their role in the improvement process.

Current standards and models for the quality of software development processes lead to a coarse-grained quality model which is heavy and difficult to focus for specific purposes. We propose a more light-weight method for assessing processes that can be expressed as process pattern languages. The method is based on imitating an existing software architecture evaluation method, ATAM, in the context of processes. The main advantages of the method are more fine-grained assessment in terms of quality attributes possibility to tune the assessment for a certain purpose, and a more light-weight assessment procedure. We illustrate the method in the case of two process pattern languages.

CMMI SCAMPI based appraisals are used worldwide to assess the process quality of organizations. In this paper we introduce a metric-based approach to assess and improve CMMI SCAMPI appraisals. To have a sound basis we at first present an appraisal meta model which defines all types of appraisal elements and their relationships. This meta model can be instantiated to get a concrete SCAMPI appraisal process, offering a precise roadmap for conducting appraisals. Based on the meta model two appraisal quality metrics are defined to systematically assess appraisal activities as well as phases and to support the improvement of appraisals. We describe the definition of these metrics in detail and give some metric interpretation guidelines.

Efficient requirements engineering and design is a demanding task. Design for excellence (DFX) offers a way to bring together different views and harmonizing practices. There are still impediments, for example, in having internal and external customers valued appropriately. The organizational implementation of DFX in itself is a debated question. We present a new way to organize the DFX concept in a large organization. The results are based on experiences of a large organization that operates in the area of ICT systems, and has had a successful implementation of the DFX concept for several years. Contrary to the traditional way of managing the DFX within R&D it is beneficial to organize it within also other parts of the operational subsystem, as this makes the concept and its improvement more visible and widespread in the organization. However, this requires seeing the concepts of problem domain and solution domain from a new angle.

Software Process Improvement (SPI) awareness is increasing among Small and Medium Sized Enterprises (SMEs). Conventional SPI frameworks are not appealing for SMEs since they are complex and costly. There are a number of frameworks, which address the problems of SMEs for SPI. This paper presents a comparative study of the most frequently referenced SPI frameworks established for SMEs from a SPI Tool coverage perspective.

Documentation is an important part of the software process, even though it is often neglected in software companies. The eternal question is how much documentation is enough. In this article, we present a practical implementation of lightweight product documentation process resulting from SPI efforts in a small company. Small companies’ financial and human resources are often limited. The documentation process described here, offers a template for creating adequate documentation consuming minimal amount of resources. The key element of the documentation process is an open source web-based bugtracking system that was customized to be used as a documentation tool. The use of the tool enables iterative and well structured documentation. The solution best serves the needs of a small company with off-the-shelf software products and striving for SPI.

IT Infrastructure Library (ITIL) is the most widely used IT service management framework that provides guidelines how to create, manage and support IT services. Service support processes, such as incident management and problem management, are among the first ITIL processes that organizations start to implement. However, several challenges may exist in the process implementation. The research question of this study is: which issues are important in establishing an ITIL-based incident management process? The main contribution of this paper is to present lessons learnt from an ITIL-based process improvement project that focused on establishing an incident management process in an IS department of a university hospital. Our results show that key issues in implementing incident management are to 1) define the basic concepts of incident management with concrete examples and 2) define process interfaces between incident management and other support processes.

Today, software-intensive systems are increasingly being developed in a globally distributed way. However, besides its benefit, global development also bears a set of risks and problems. One critical factor for successful project management of distributed software development is the allocation of tasks to sites, as this is assumed to have a major influence on the benefits and risks. We introduce a model that aims at improving management processes in globally distributed projects by giving decision support for task allocation that systematically regards multiple criteria. The criteria and causal relationships were identified in a literature study and refined in a qualitative interview study. The model uses existing approaches from distributed systems and statistical modeling. The article gives an overview of the problem and related work, introduces the empirical and theoretical foundations of the model, and shows the use of the model in an example scenario.

The market is becoming more and more competitive, a lot of products and services depend of the software product and the software is one of the most important assets, which influence the organizations’ businesses. Considering this context, we can observe that the companies must to deal with the software, developing or acquiring, carefully. One of the perspectives that can help to take advantage of the software, supporting effectively the business, is to invest on the organization’s software processes. This paper presents an approach to evaluate and improve the processes assets of the software organizations, based on internationally well-known standards and process models. This approach is supported by automated tools from the TABA Workstation and is part of a wider improvement strategy constituted of three layers (organizational layer, process execution layer and external entity layer). Moreover, this paper presents the experience of use and their results.

Value-based software engineering (VBSE) is an emerging stream of research that addresses the value considerations of software and extends the traditional scope of software engineering from technical issues to business-relevant decision problems. While the concept of value in VBSE relies on the well-established economic value concept, the exact definition for this key concept within VBSE domain is still not well defined or agreed upon. We argue the discourse on value can significantly benefit from drawing from research in management, particularly software business. In this paper, we present three aspects of software: as a technology, as a design, and as an artifact. Furthermore, we divide the value concept into three components that are relevant for software product development companies and their customers: intrinsic value, externalities and option value. Finally, we propose a value decomposition matrix based on technology views and value components.

Industry managers have long recognized the vital importance of information security for their businesses, but at the same time they perceived security as a technology-driven rather then a business-driven field. Today, this notion is changing and security management is shifting from technology- to business-oriented approaches. Whereas there is evidence of this shift in the literature, this paper argues that security standards and academic work have not yet taken it fully into account. We examine whether this disconnect has lead to a misalignment of IT security requirements in businesses versus industry standards and academic research. We conducted 13 interviews with practitioners from 9 different firms to investigate this question. The results present evidence for a significant gap between security requirements in industry standards and actually reported security vulnerabilities. We further find mismatches between the prioritization of security factors in businesses, standards and real-world threats. We conclude that security in companies serves the business need of protecting information availability to keep the business running at all times.

Waterfall development is still a widely used way of working in software development companies. Many problems have been reported related to the model. Commonly accepted problems are for example to cope with change and that defects all too often are detected too late in the software development process. However, many of the problems mentioned in literature are based on beliefs and experiences, and not on empirical evidence. To address this research gap, we compare the problems in literature with the results of a case study at Ericsson AB in Sweden, investigating issues in the waterfall model. The case study aims at validating or contradicting the beliefs of what the problems are in waterfall development through empirical research.

The amount of software is increasing in the different domains in Europe. This provides the industries in smaller countries good opportunities to work in the international markets. Success in the global markets however demands the rapid production of high quality, error free software. Both CMMI and agile methods seem to provide a ready solution for quality and lead time improvements. There is not, however, much empirical evidence available either about 1) how the integration of these two aspects can be done in practice or 2) what it actually demands from assessors and software process improvement groups. The goal of this paper is to increase the understanding of CMMI and agile integration, in particular, focusing on the research question: how to use ‘lightweight’ style of CMMI assessments in agile contexts. This is done via four case studies in which assessments were conducted using the goals of CMMI integrated project management and collaboration and coordination with relevant stakeholder process areas and practices from XP and Scrum. The study shows that the use of agile practices may support the fulfilment of the goals of CMMI process areas but there are still many challenges for the agile teams to be solved within the continuous improvement programs. It also identifies practical advices to the assessors and improvement groups to take into consideration when conducting assessment in the context of agile software development.

Enterprise Resource Planning (ERP) systems provide integration and optimization of various business processes which leads to improved planning and decision quality, smoother coordination between business units resulting in higher efficiency, and quicker response time to customer demands and inquiries. This paper reports challenges, opportunities and outcome of ERP implementation in Oil & Gas exploration sector. This study will facilitate in understanding transition, constraints and implementation of ERP in this sector and also provide guidelines from lessons learned in this regard.

Software is one of the most important drivers of innovation. As organizations are becoming more dependent on software, the improvement of software quality and productivity becomes of essential importance for the competitiveness of an organization. Continuing the success of the LSO Workshop series since 1999, this workshop will provide a communication forum bringing together academia and industry for discussing the advancements made and addressing the challenges faced by continuous learning in software-intensive organizations.

Software companies who want to improve software process capabilities (SPCs)a systematic method to make informed investment decisions on software process improvement (SPI) initiatives. Such decisions should aim at creating maximum stakeholder values. To address this problem, we present a method with tool support that may help companies align stakeholder values with SPCs and SPI initiatives. The proposed method has been developed based on the well-established “Quality Function Deployment” (QFD) approach. The experience with the proposed method suggests that it particularly helps to reduce the risk of misalignment by identifying those SPI initiatives that are most beneficial to stakeholders. The tool support provided with the proposed method also generated positive experiences in increasing the usability of the method and helped companies in the elicitation and prioritization of stakeholder values. Therefore, we propose a workshop for the method work out named “Smarter Investment by Aligning SPI Initiatives, Capabilities and Stakeholder Values” in hypothetical case company.

Most of today’s products and services are software-based. Organizations that develop software want to maintain and improve their competitiveness by controlling software-related risks. To do this, they need to align their business goals with software development strategies and translate them into quantitative project management. There is also an increasing need to justify cost and resources for software and system development and other IT services by demonstrating their impact on an organization’s higher-level goals. For both, linking business goals and software-related efforts in an organization is necessary. However, this is a challenging task, and there is a lack of methods addressing this gap.

Working in the globally distributed market is one of the key trends among the software organizations all over the world. [1-5]. Several factors have contributed to the growth of distributed software development; time-zone independent ”follow the sun” development, access to well-educated labour, maturation of the technical infrastructure and reduced costs are some of the most commonly cited benefits of distributed development [3, 6-8]. Furthermore, customers are often located in different countries because of the companies’ internationalization purposes or good market opportunities.

This document presents a tutorial on case study research methodology in software engineering, held at the 10th International Conference on Product Focused Software Development and Process Improvement (Profes).