Index Data's DPLA submission

Index Data has posted a statement of interest to the DPLA (Digital Public Library of America) beta sprint. We are submitting this together with two academic partners.

The proposal in many ways reflects our broader thinking about information architectures and the flow of metadata between organizations, so we thought it would be useful to share it here. Fans of ID may notice that, yes, we are looking at Linked Data too. This will evolve into a detailed proposal between now and September. Users of our tools as well as others who like open discovery technologies are welcome to get in touch if you’re interested in discussing this with us.

This proposal addresses the issue of search and discovery in DPLA – in particular, the integration of many different information sources into one coherent framework that is capable of organic growth as technology, practices, and end-user expectations evolve.

Any coherent digital library architecture will depend on an easy flow of metadata between systems: From the simplest ‘catalog’ interface to tools supporting the digitization process, such as collection analysis; prioritizing and scheduling scans, and copyright status analysis. Increasingly, systems can be expected to draw on Linked Data models, since bibliographic data is inherently rich in relations and references that are only poorly represented by current practices. The types of relationships that can be more precisely expressed by Linked Data models include information about authors (How do you know you’re talking about the same author), their creative works, and the various expressions and manifestations of these.

Metadata will be used in many scenarios – by end-users as well as the catalogers and curators of the digital library. However, even for end-users, we believe that rather than just a single, public search box, the DPLA must have many faces in many different contexts – from localized library websites to applications for mobile devices and social media sites and applications as yet unthought of. The essence of the individual library is its unique relationship with its local community, so the DPLA architecture must support the notion of many individual organizations each acting independently and participating in its own way, coloring its view of the whole to suit the requirements of its constituents.

The best way to meet these requirements is to propose a metadata architecture for the DPLA based on open specifications and open source software to support its initial implementation. Open interfaces and a service-oriented architecture reduce the risk of monolithic software solutions or dependencies on particular computing platforms or development environments: They stimulate an open marketplace of ideas.

The architecture will make use of numerous existing technologies, including, SRU, OAI-PMH, and existing and emerging data models such as DublinCore and RDA. The reference implementation will use open source software released both by ourselves and by others. It will include example user interfaces.

Index Data specializes in interoperability around metadata through open standards and open source software. We have collaborated with academic organizations, libraries, and government bodies. The partnership behind this proposal has a rich history of research and practical development in the field.