Catálogo de publicaciones - libros

The amount of software in many electronic products is growing rapidly. Two examples: the amount of software in a mobile phone is expected to increase from 2 million today to 20 million in 2010. The amount of software in a car in 2010 is expected to be 100 million lines of source code (Charrette, 2005). Many companies see their business change from a hardware business to a software business. The impact on companies goes far beyond development and development processes. Adoption of proper software sales and legal processes is as important. The presentation is based on research into the software business and 10 years of experience in managing software companies.

In 2000, the Software Engineering Institute introduced the continuous representation of the CMMI. In the following years, many organizations based their process improvement effort on this representation. Despite the advantages of this model, several of these organizations found it hard to make a choice of which process areas to implement first. To help organizations make these decisions, three improvement paths are recognized: Project, Process and Product. This so-called PPP concept of continuous and the improvement paths will be addressed in this keynote presentation.

This keynote addresses the evidence-based paradigm currently being adopted in many practical sciences (e.g., medicine, education, social policy) and discusses whether it is applicable to software engineering. In the presentation, the view is taken that although Evidence-based Software Engineering may be unproven, one aspect of the evidencebased paradigm is hard to ignore, that is: Systematic literature reviews. Systematic literature reviews aim to summarize research studies related to a specific research question in a way that is fair, rigorous, and auditable. The keynote presentation will outline the potential benefit of systematic literature reviews and describe in detail the process of performing such a systematic literature review.

Software product families have found broad adoption in Nokia, the telecom industry, and the embedded systems industry as a whole. Product family thinking has been prevalent in this context for mechanics and hardware, and adopting the same for software has been viewed as a logical approach. During recent years, however, the trends of convergence, end-to-end solutions, shortened innovation and R&D cycles and differentiation through software engineering capabilities have led to a development where organizations are stretching the scope of their product families far beyond the initial design. Failing to adjust the product family approach, including the architectural and process dimensions, when the business strategy is changing is leading to several challenging problems that can be viewed as symptoms of this approach. The presentation discusses the key symptoms, the underlying causes for these symptoms as well as solutions for realigning the product family approach with the business strategy. The presentation uses examples from Nokia to illustrate the solutions and approaches that will be discussed.

This paper outlines a process for software system modernization decisions. The rationale of the process is explained and the process is defined in a way that allows its adaptation for other organizations and situations. The process is a light-weight one and is based on the use of objective data. The procedures for collecting the data are explained. The process has been used to solve a real industrial decision making situation in which the process was successful.

The release planning activity in market-driven requirements engineering is crucial but difficult. The quality of the decisions on product content and release timing determines the market success, but as predictions of market value and development cost are uncertain, the decisions are not always optimal. This paper presents a prototype tool for retrospective analysis of release planning decisions based on tool requirements gathered in two previous empirical studies where retrospective analysis was done manually. The supported method enables representation of different views in the decision-making process. The results from an initial validation indicate that the retrospective analysis benefits from the supporting tool.

The web plays a central role in such diverse application domains as business. As the use of web grows, organizations are increasingly choosing to use it to provide their services to their clients. Services are the systemization of the business processes in the organization. A bad definition and management of the services makes the systematization fail or not to have the expected success. The business process modeling is the first step in the systematization. Due to the great number of modeling tools in existence it is necessary to identify the information that they allow to specify. In this paper, a set of concepts is proposed to evaluate modeling tools for business process modeling using three levels of abstractions –organizational, conceptual and web. The evaluation compares the modeling capabilities supplied by the different techniques. This evaluation also allows determining what modeling tool is the most appropriate to model specific concepts of interest to a particular organization or problem.

This paper discusses about the improvement of source code review process for embedded software development project, and also proposes an effective approach to source code review for embedded software. As the start point of the discussion, this paper firstly discusses the results of a survey we conducted of about 290 embedded software development projects in Japan from the viewpoint of quality. Next, this paper discusses the problems of current source code review process and the way for improvement of the process. In the discussion, we focus on quality characteristics in ISO/IEC9126 and apply this to our improved review process. That is, we propose a new review process which is based on selection of review target portion in the target source code and selection of the review check items. As for the selection of review check items, using the characteristics viewpoints in ISO/IEC9126, review check items are selected according to the target software features.

Many modern new product development (NPD) embedded software projects are required to be run under turbulent conditions. Both the business and the technological environments are often volatile. Uncertainty is then an inherent part of the project management. In such cases, traditional detailed up-front planning with supporting risk management is often inadequate, and more adaptive project management tools are needed. This industrial paper investigates the typical problem space of those embedded software projects. Based on a literature survey coupled with our practical experiences, we compose an extensive structured matrix of different potential project problem factors, and propose a method for assessing the project’s problem profile with the matrix. The project manager can then utilize that information for problem-conscious project management. Some industrial case examples of telecommunications products embedded software development are illustrated.

Besides the promise of rapid and efficient software development, agile methods are well-appreciated for boosting communication and motivation of development teams. However, they are not practical “as such” in large organizations, especially because of the well-established, rigid processes in the organizations. In this paper, we present a case study where a few agile practices were injected into the software process of a large organization in order to pilot pair programming and improve the motivation and competence build-up. The selected agile practices were pair programming, the planning game and collective code ownership. We show how we adjust these practices in order to integrate them into the existing software process of the company in the context of a real software project.