Version Control of structured data: A study on different approaches in XML

Abstract

The structured authoring environment has been changing towards a decentralised form of authoring. Existing version control systems do not handle these documents adequately, making it very difficult to have parallel authoring in a structured environment. This study attempts to find better alternatives to the existing paradigms and tools for versioning XML documents. To achieve this, the DESMET methodology for evaluating software engineering methods and tools was applied, with both a Feature Analysis and a Benchmark Analysis being performed. Concerning the feature analysis, the results demonstrate that the XML-aware tools are, as expected, better at XML specific concerns, such as considering the history of a specific node. Conversely, the non-XML-aware ones are not able to achieve good results in the XML specific concerns, but do achieve a high score when considering project maturity or general repository management features. Regarding performance, this study concludes that XML-aware tools bring a considerable overhead when compared to the non-XML-aware tools. This study concludes that the selection of an approach to versioning XML should be dependent of the priorities of the documentation project.