National Boundaries and Semantics of Artefacts in Open Source Development

Global software development has long being recognised as a paradigm shift in modern software development. As an immediate effect, co-location of workers in the same building or office is not seen as necessary any longer. Coordination in distributed socio-technical systems is mostly achieved by means of the artifacts that are produced by the developers part of a project's team. Geographic distance profoundly affects the ability to collaborate. With communication becoming less frequent, the challenge is for it to become more effective. This is especially complex when different nationalities, languages and cultures are part of the same development effort. Open source software is an example of a distributed, multi-lingual development effort. As such, the main resulting artefacts are discussions, and source code. Diverse backgrounds can produce a different semantic corpus if the authors come from the same ethnic and language groups or from different ones. The purpose of this paper is to evaluate the artifacts in the context of their semantics, and how semantic corpora are affected by development and languages. By using a selection of Open Source projects developed within national boundaries, we compare their semantic richness, and how their class content is reflected in their identifiers. We also compare these national projects to a successful, international project. The aim is to discover how national boundaries influence the semantics of the developed code.