Quality-oriented software architecture development

Producing software products of good quality requires that quality requirements are taken into account as early as possible. In theory, the first place in which quality requirements can be addressed is architectural models of software. However, in practice, the software’s architecture is only used to describe the functionality of the developed software. This means that the implemented software may not fulfil its quality requirements and some parts of the implementation process might be useless. The main research problem in this work is how to define and connect quality requirements with the software architecture in such a way that the requirements can vary between software components and software family members. An environment for defining and collecting software’s quality requirements was designed and implemented in this work. The environment consists of three main parts: quality meta-data management, quality modelling and quality evaluation. The quality meta-data management provides a possibility to define each quality attribute into an ontology form. These ontologies are utilized when quality requirements are defined in order to define the requirements in a uniform way in the quality modelling phase. Quality requirements are defined according to the UML profile developed for that purpose, so that it would be possible to represent these requirements in the architectural models. Finally, the architecture’s quality is evaluated using evaluation tools. The purpose was to implement the whole environment on the Eclipse platform using available open source components. The Eclipse was selected because it is a widely used open source platform, which makes it easier to distribute the software developed during this work. The Eclipse tool evaluation confirmed that TOPCASED is the best available UML tool for the Eclipse and the best Eclipse ontology tool is EODM. However, EODM does not fulfil all the desired requirements of this work and this enforced the use of the Protégé ontology tool.