Architectural Properties for Data Reusability

A fundamental requirement for long-term digital repositories is to ensure the reusability of data that have been entrusted to data stewardship. This is especially true for scientific studies involving repeated observations of the same variables during different time periods. The data must be preserved for future use and remain usable, even as science discipline practices change and implementation technologies evolve. The Planetary Data System (PDS) has about four decades of experience building a digital repository for diverse and evolving science domains. The PDS has been successfully providing its designated community with high-quality peer-reviewed digital data that are being reused regularly in scientific studies, cross-disciplinary research, and data science. The PDS4 Information Model was designed using principles that promote data reusability in long-term digital repositories. A key design principle was that a “single shared ontology” is critical for developing a digital repository to enable semantic interoperability across disciplines. Also used were international standards for long-term digital preservation and metadata management and a process of knowledge acquisition involving science domain experts. Several architectural properties have been identified as important for promoting data reusability.