Enhancement of a software chrestomathy for open linked data

A software chrestomathy collects small software systems as an aid in learning a subject. As many languages, technologies and concepts might be involved, it will contain a heterogeneous mix of code artifacts, documentation and relationships. This is problematic, as knowledge resources should convey their data in a structured manner. The data should be conveniently explorable, easily discoverable and as accessible for humans as well as machines. This thesis tries to tackle the problems created by the heterogeneousness of the data by applying Linked Data principles. The 101companies chrestomathy is enriched with these principles, meaning that every important entity is seen as a resource and dereferencable through HTTP, which results in meaningful data about this entity. Additionally, the entities are linked with each other, so that all available data is reachable. It is shown that by embracing the Linked Data principles, the problems created by the diverse data sets can be alleviated. Furthermore, examples are given on how the approach can even enable further research options, such as clone detection, that were previously difficult if not unfeasible. Acknowledgements The 101companies system is a joint work of the Softlang Team at the University of Koblenz-Landau. At the time of writing, it relies on contributions of Kevin Klein, Aleksey Lashin, Ralf Lämmel, Arkadi Schmidt, Thomas Schmorleiz and Andrei Varanovich. This thesis would not have been possible without them.