Marcus Zinn


The reuse of software units, such as classes, components and services require professional knowledge to be performed. Today a multiplicity of different software unit technologies, supporting tools, and related activities used in reuse processes exist. Each of these relevant reuse elements may also include a high number of variations and may differ in the level and quality of necessary reuse knowledge. In such an environment of increasing variations and, therefore, an increasing need for knowledge, software engineers must obtain such knowledge to be able to perform software unit reuse activities. Today many different reuse activities exist for a software unit. Some typical knowledge intensive activities are: transformation, integration, and deployment. In addition to the problem of the amount of knowledge required for such activities, other difficulties also exist. The global industrial environment makes it challenging to identify sources of, and access to, knowledge. Typically, such sources (e.g., repositories) are made to search and retrieve information about software unitsand not about the required reuse activity knowledge for a special unit. Additionally, the knowledge has to be learned by inexperienced software engineers and, therefore, to be interpreted. This interpretation may lead to variations in the reuse result and can differ from the estimated result of the knowledge creator. This makes it difficult to exchange knowledge between software engineers or global teams. Additionally, the reuse results of reuse activities have to be repeatable and sustainable. In such a scenario, the knowledge about software reuse activities has to be exchanged without the above mentioned problems by an inexperienced software engineer. The literature shows a lack of techniques to store and subsequently distribute relevant reuse activity knowledge among software engineers. The central aim of this thesis is to enable inexperienced software engineers to use knowledge required to perform reuse activities without experiencing the aforementioned problems. The reuse activities: transformation, integration, and deployment, have been selected as the foundation for the research. Based on the construction level of handling a software unit, these activities are called Software Construction Activities (SCAcs) throughout the research. To achieve the aim, specialised software construction activity models have been created and combined with an abstract software unit model. As a result, different SCAc knowledge is described and combined with different software unit artefacts needed by the SCAcs. Additionally, the management (e.g., the execution of an SCAc) will be provided in a service-oriented environment. Because of the focus on reuse activities, an approach which avoids changing the knowledge level of software engineers and the abstraction view on software units and activities, the object of the investigation differs from other approaches which aim to solve the insufficient reuse activity knowledge problem. The research devised novel abstraction models to describe SCAcs as knowledge models related to the relevant information of software units. The models and the focused environment have been created using standard technologies. As a result, these were realised easily in a real world environment. Softwareengineers were able to perform single SCAcs without having previously acquired the necessary knowledge. The risk of failing reuse decreases because single activities can be performed. The analysis of the research results is based on a case study. An example of a reuse environmenthas been created and tested in a case study to prove the operational capability of the approach. The main result of the research is a proven concept enabling inexperienced software engineers to reuse software units by reusing SCAcs. The research shows the reduction in time for reuse and a decrease of learning effort is significant.

Document Type


Publication Date