Automatic Discovery of Preservation Alternatives Supported by Community Maintained Knowledge Bases

Preservation Planning, which deals with selecting the most appropriate preservation action to be applied to digital objects, is an important step in any digital preservation activity. Comprehensive Preservation Planning depends on the availability of identified alternatives of preservation actions, which are for example file format migrations to migrate data in an outdated format to one that has better support. Also emulation, e.g. of the behaviour of a specific software application (application emulation), can be a viable preservation action. The alternative identification step can either be performed manually by an expert, or (semi-)automatically, if appropriate knowledge bases are available. Building and maintaining such knowledge bases is however a tedious task, as the number of software applications and file formats, and especially their relation to each other, is very large. In this paper, we therefore present an approach to automatically build knowledge bases for Preservation Planing from already existing, open resources. One such source is the community maintained Freebase, which contains linked data on many topics, among them file formats, software applications, and most importantly, their relations, in a structured manner. We demonstrate the applicability of these knowledge bases by automatically identifying possible digital preservative actions on a uses case, an eScience experiment from the domain of data mining. This use case originates from the task of process preservation, where we look beyond single files, but regard complete chains of executions as the objects to be preserved.