Triple Bypass – Open MARC Surgery
As more and more datasets become part of the semantic web, originating both within and outside the library community, actually consuming and using Linked Data becomes increasingly relevant. As a part of building a new version of Libris, the Swedish Union Catalogue, Linked Data technologies are now an integral part of the core of the system. This has brought about a radical change in how we do systems design making the distinction between internal and external datasets blurry at best. It also raises a couple of questions: how do we deal with changes in datasets outside our control, how do we tell people that data they relate to has changed? And ultimately: what about MARC? As a part of a larger library infrastructure dealing with MARC-records is not optional, but it does not have to be at the centre of the system. And it certainly does not need to have a large impact on the design.
Ultimately, building a system that is “linked by default” provides something that the older systems could not: context. Context in turn is what needed if we want to bring the library to the user. Being able to draw from data from other cultural institutions, Wikipedia, etc. makes it easier, or rather possible, to create compelling interfaces for our users. This, is the driving force for Linked Data when we move beyond simply publishing it.
This presentation details the systems design decisions made while building a union catalogue and cataloguing interface on Linked Data rather than traditional MARC-records as well as the benefits and examples of the same.
Martin Malmsten is Head of Software Development, Libris