Navegación

Búsqueda

Búsqueda avanzada

An industrial case study on the use of UML in software maintenance and its perceived benefits and hurdles (RELEVANTE YA PUBLICADO)

UML is a commonly-used graphical language for the modelling ofsoftware. Works regarding UML’s effectiveness have studied projects thatdevelop software systems from scratch. Yet the maintenance of softwareconsumes a large share of the overall time and effort required to develop softwaresystems. This study, therefore, focuses on the use of UML in softwaremaintenance. We wish to elicit the practices of the software modelling usedduring maintenance in industry and understand what are perceived as hurdles andbenefits when using modelling. In order to achieve a high level of realism, weperformed a case study in a multinational company’s ICT department. Theanalysis is based on 31 interviews with employees who work on softwaremaintenance projects. The interviewees played different roles and providedcomplementary views about the use, hurdles and benefits of software modellingand the use of UML. Our study uncovered a broad range of modelling-relatedpractices, which are presented in a theoretical framework that illustrates howthese practices are linked to the specific goals and context of softwareengineering projects. We present a list of recommended practices that contributeto the increased effectiveness of software modelling. The use of softwaremodelling notations (like UML) is considered beneficial for softwaremaintenance but needs to be tailored to its context. Various practices thatcontribute to the effective use of modelling are commonly overlooked,suggesting that a more conscious holistic approach with which to integratemodelling practices into the overall software engineering approach is required.

An extension to Simply for solving Weighted Constraint Satisfaction Problems with Pseudo-Boolean Constraints

Max-Simply is a high-level programming framework for modelling and solving weighted CSP. Max-Simply can also deal with meta-constraints, that is, constraints on constraints. The technology currently used to solve the generated problem instances is SMT. In this paper we present a variant of Max-Simply which is able to generate not only SMT instances but also pseudo-Boolean instances for certain modellings. Since there are problems that are more naturally encoded using pseudo-Boolean variables, the possibility of generating pseudo-Boolean instances can result in a more efficient and natural fit in some situations. We illustrate the expressiveness of the Max-Simply language by modelling some problems, and provide promising performance results on the corresponding generated pseudo-Boolean instances using state-of-the-art pseudo-Boolean solvers.