Catálogo de publicaciones - libros

The adoption of a Software Product Line approach implies a series of changes in the way an organization develops software and runs its whole business. This change in the organization’s business strategy can lead to knowledge gaps between the knowledge the organization has at present and the knowledge it must have in the future in order to implement its new strategy. In this article we propose to consider the transition to a Product Lines approach as a Knowledge Management problem, and we also introduce a method for identifying and assessing the aforementioned knowledge gaps.

A reference model for the comparison of system family modeling approaches is presented. Three main approaches to system family modeling are illustrated with a simple example and compared relative to the reference model.

Economic models for reuse are very important to organizations aiming to develop software with large scale reuse approaches. In fact, the initial investment is so important that it can discourage managers to commit to those approaches. Thus, economic models can help them to assess the worthiness of such an investment. Product Line Engineering (PLE) seems to be an attractive reuse approach in matter of product quality and time-to-market. Using Commercial Off The Shelf (COTS) in a PLE approach may have a positive impact. This paper reports on the need for an economic model to quantify the predicted benefits of the PLE software development with the use of COTS components. We introduce a Model for Software Cost Estimation in a Product Line Engineering approach that we denote SoCoEMo-PLE 2. This latter includes the usage of COTS components. The potential benefits of the model are described.

Active innovation management is performed by companies to create an environment that fosters innovation. In a product line environment, platform and predefined variability restrict innovation because the development artifacts in the platform and the variation are prescribed. An analysis of innovation projects in literature shows that moderate innovations like introducing a new member of a product line yield only a small return on investment. This paper introduces a series of measures that can help to prevent a lock-in of a product line organization with respect to innovation. We take a look at various aspects of innovation – personnel, customer and market, technology and engineering, organization and process. Organizations may pick the best-suited measures for their current situation.

There is ample evidence that for many software development organizations, a change to software product line practice would be good . But there is also a reluctance by many of these software development organizations to go first . This panel explores what remains to be done by the SPLC community and others to overcome the inhibitors and to facilitate SPL adoption.

Proponents of different software product line and software reuse economic models will be given a real-world software product line scenario and asked to predict alternate outcomes and to justify – with hard data from their models – some of the difficult choices that need to be made in the scenario. The audience will have a chance to compare, side by side, the predictions, recommendations, insights, intuitive fidelity and ease-of-use of the different models.

One of the remaining challenges in product line engineering is how to establish the quality of the reusable assets so that we can be confident that they can be configured and composed reliably. This is desirable, both to avoid having to completely re-test each product and to avoid integration faults only being detected late in product development. One of the diversity mechanisms of Philips’ high-end TV product line is the selection and com position of sub-systems, so different sub-system variants must integrate reli ably if the aims of the product line are to be realized. An earlier study of in tegration testing obligations in Philips products concluded that certain design policies must be imposed if integration testing is to be feasible, but it did not describe how relevant policies could be identified at the earliest stages of de sign. This paper addresses how a set of architectural rules were established for the TV product line through a root-cause analysis of problem reports, and packaged so that developers can recognize when they should be applied. The approach builds on other work on the impact of design choices on non-func-tional requirements to ensure that all quality attributes are addressed.

Our society is becoming increasingly dependent on embedded software, and its reliability becomes more and more important. Although we can utilize powerful scientific methods such as model checking techniques to develop reliable embedded software, it is expensive to apply these methods to consumer embedded software development. In this paper, we propose an application of model checking techniques for design verification in product line development (PLD). We introduce reusable verification models in which we define variation points, and we show how to define traceability among feature models, design models and verification models. The reuse of verification models in PLD not only enables the systematic design verification of each product but also reduces the cost of applying model checking techniques.

In today’s competitive business environment, it is extremely important to offer customers exactly the products they want. Software product lines have the potential to enable companies to offer a large variety of products while still being able to manage the complexity caused by this increased number of products. But offering a large range of variants does not necessarily mean increased profits, as many manufacturing companies had to notice in the early 1990ies. The task of Product Portfolio Planning is the development of a product portfolio that optimally satisfies customer demands and at the same time restricts the number of products offered. Quality Function Deployment (QFD) is a well-known and successfully used Quality Management method that can help companies to identify true customer needs and the features needed to fulfil these needs. This paper demonstrates how QFD can be used for Product Portfolio Planning, thus offering potentially great benefits.

In the product-line context, where a lack or mismatch in a quality attribute is potentially replicated among all products, product-line evaluation could detect problems before concrete products are developed. The life span of a software product-line architecture is much longer than the one of an ordinary software product and it serves as a basis for a set of related systems. Therefore, the product-line architecture should be adaptable to evolution as well as support a number of different products. All these characteristics set new requirements to the product-line architecture evaluation. This paper highlights the new issues that can arise when evaluating a product-line architecture versus evaluating a single-system architecture, including classifications of relevant attributes in product-line architecture evaluation, new evaluation moments and techniques. These issues are used as components of a framework to survey product-line architecture evaluation methods and metrics.