Uma abordagem de Data Warehouse para análise de processos de desenvolvimento de software

• Ano de defesa: 2006
• Autor(a) principal: Novello, Taisa Carla
• Orientador(a): Becker, Karin
• 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; BANCO DE DADOS - ARQUITETURA; DATA WAREHOUSE
• Link de acesso: http://hdl.handle.net/10923/1570

Software quality is important and necessary for organizations. They are interested on how to measure and quantitative analyze the processes quality. As organizations work with different projects, processes and metrics, they should looking for alternatives that provide a unified vision using the centralization of different projects data and provide quantitative analysis about Software Development Processes (SDP) through a Metric Program (MP). Thus, the software maturity models suggest the construction of a organizational repository of metrics. However, the construction of a repository that attends the data management and analysis functionality is not a trivial task. This work describes briefly the environment architecture for Data Warehousing that provides support for MP adoption through data stored that means different processes in a unified database. This volume describes two components: the analytic model based on the Data Warehouse (DW), and the presentation component that uses analytic resources to facilitate the users analyses. The development of a repository should consider the particularities of MP and the environment of PDS. About the metrics that compose MP, some of them represent non addictive data what can made the analysis difficult about the environment, the PDS features made difficult the unified model definition that supports distinct characteristics. Besides the data storage, the manner who it is available also must be considered once the users has different analysis needs. Thus, the complexity of the model development and analysis resource availability is high. Thus, the analytic model proposed intends to store metrics and SDP data, according to analysis requirements, considering the MP features particularity and the SDP environment. The definition of analytic resources considers users with different profiles. These resources attend the analysis requirements for each profile and shows the information through many granularity levels, including mechanisms that provide more semantic for the data. Thus, this work provides an infra-structure that supports different SDP and quantitative analyses for different profiles base on MP.