No abstract

Industrial evaluations of COTS software mostly use the quality models provided by international standards. But the context and objectives of COTS evaluations are fundamentally different from those primarily defined by the standards. Some key issues are often forgotten: 1°) the existence of several evaluators and several quality models sharing common factors, criteria and measures, 2°) the purpose of the evaluation model, 3°) measures of different types, and 4°) the recursive nature of the model since each node is an evaluation model itself. We had the occasion to study the results of real standard-based COTS evaluations. Faced with the difficulties of exploiting them, we experimented the use of multi-criteria methodology. This work allows us to understand some of the problems generated by applying the standards to COTS evaluations, and to propose new principles for evaluating software quality that should be considered in an evolution of the standards. This paper reports our experiment.

RésuméNous présentons, dans ce papier, un support d'aide à la conception collaborative de produits appartenant à une famille de produits. Le produit est modélisé par un arbre de composants et les contraintes de développement, utilisées pour guider le processus de conception, sont exprimées à l'aide d'un langage spécifique basé sur des expressions logiques. L'objectif principal de ProjectLeader est d'économiser le temps de conception en favorisant le travail simultané et en prévenant les retours en arrière. Ainsi, pour chaque intention de décision émise par un concepteur, ProjectLeader vérifie que la décision ne remettra pas en cause la possibilité d'atteindre un état correct du produit final, c'est-à-dire un état dans lequel toutes les contraintes sont vérifiées. Mots-clefs : conception collaborative, système à base de composants, modélisation et traitement de contraintes de conception, modélisation et traitement de la variabilité AbstractIn this paper we present a support that helps organizing distributed design of products belonging to a product family. We model the product to be developed as a component-based tree of object templates, develop a logic-based language that allows expressing diverse development constraints and use these constraints to guide the design process. The main objective of our support is to save time and development effort in increasing parallel work and avoiding roll-backs. For that, it reasons in terms of correctness of the future, final state of the product, and verifies that the product state after each operation allows the reacheability of such a state.