Uma abordagem orientada a serviços para captura de métricas de processo de desenvolvimento de software

• Ano de defesa: 2006
• Autor(a) principal: Cunha, Virginia Silva da
• Orientador(a): Ruiz, Duncan Dubugras Alcoba
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Porto Alegre
• Palavras-chave em Português: INFORMÁTICA; ENGENHARIA DE SOFTWARE; DATA WAREHOUSE; BANCO DE DADOS
• Link de acesso: http://hdl.handle.net/10923/1639

Software organizations work with several software projects that differ in terms of both the management tools used and the way tracking metrics are stored and controlled. Thus, the lack of a central data repository poses difficulties for the traking of Software Development Processes (SDPs) in these organizations. One of the crucial steps of the Knowledge Discovery in Databases Process is the process of Extraction, Transformation and Load (ETL). ETL aims to transform the raw data extracted from different fonts into consistent, reliable information. Considering the SDPs specificities, this study was carried out in the real computional environment. It was observed that the tools used range from spreadsheets (e. g. MS Excel) to control tools for the execution of project activities (e. g. MS Project Server, IBM Rational Clear Quest, Bugzilla). Different SDP models with a distinct life cycle for each project are also used, which result in completely different ways to register these projects even when using the same tool. Another problem is that each of those tools has an own data model that does not follow defined data representation standards. Therefore, the extraction of those data becomes a challenging goal to achieve, raising the complexity of ETL processes. The model proposed in this study introduces a two-integrated approach to deal with the problem: 1) a non intrusive way of data extraction, taking several types of heterogeneities into account, 2) the transformation and integration of these data, providing a unified and quantified organizational view of the projects. These aspects are treated using a serviceoriented architecture. This service oriented architecture tries to deal with several types of heterogeneity, from both the technical (e. g. different tools) and organizational standpoint (e. g. Organization’s Standard Software Process Standard specializations that result in distinct ways to develop and register project facts). This heterogeneity is conveniently treated through services that work as wrappers of the different types of extractors and through the support of a distributed development environment. For the evaluation of the proposed approach, three examples that consider all heterogeneity issues (different types of projects, different life cycles, different management models and several management support tools) were developed.