An Associative Documentation Model

This diploma work examines the MAINTENANCE OF TECHNICAL DOCUMENTATION within an software engineering process. The characteristics of technical documentation and its behaviour within an dynamic software development environment are important to understand the problems that occur with technical documentation. I explore the factors that influence the development and the resulting quality of the technical documentation. The RELATIONSHIP BETWEEN SOFTWARE SOURCE CODE AND TECHNICAL DOCUMENTATION is used to coordinate the development of the technical documentation with the software development. The principle to match software entity names with documentation segments defines the relationship between software and documentation. I demonstrate how it works and how it is used for coordination of the software development and the technical documentation development. An analysis of different name representations and documentation segmentation structures shows the influence of the structures on the creation of relations. I explore under which conditions relations are generated that fit best to the relationship between software and documentation that exists in reality. The ASSOCIATIVE DOCUMENTATION MODEL (ADM) builds on the relationship between software and documentation that is determined by matching of software entity names within documentation. The ADM focuses on three aspects: It concentrates on the extraction of the names of software entities. ADM uses Famix models that are capable to represent any object-oriented software and detects the software entities to be represented by their names. Additionally, ADM considers structural relationships between the software entities that are given by inheritance, aggregation, invocation and access. The second aspect is the representation of the relationship between software and documentation. Especially the influence of the inner relationship of software entities on the relationship between software and documentation is important. The third aspect is to get the detection of the software entity names and the generation of the relationship between software and documentation into a consistent model. It serves as a uniform model of the software-documentation relationship for any application that uses this model. CHANGE IMPACT DETECTION is an sample application of the ADM. It determines changes between two software versions by comparison of the models of these versions. Differences between the models are interpreted as changes. The names of the entities that are affected by the changes are taken as representations of the changes. The ADM relates these change representations to the documentation. This way, impact of the software changes on the documentation is detected over the ADM relations. The functionality and usage as well as the power and limitations of Change Impact Detection