Forming Consensus on Schema.org for Libraries and More

Back in September I formed a W3C Group – Schema Bib Extend.  To quote an old friend of mine “Why did you go and do that then?” 

Well, as I have mentioned before Schema.org has become a bit of a success story for structured data on the web.  I would have no hesitation in recommending it as a starting point for anyone, in any sector, wanting to share structured data on the web.  This is what OCLC did in the initial exercise to publish the 270+ million resources in WorldCat.org as Linked Data.

At the same time, I believe that summer 2012 was a bit of a watershed for Linked Data in the library world.  Over the preceding few years we have had various national libraries publishing linked data (British Library, Bibliothèque nationale de France, Deutsche National Bibliothek, National Library of Sweden, to name just a few).  We have had linked data published versions of authority files such as LCSH, RAMEAU, National Diet Library, plus OCLC hosted services such as VIAF, FAST, and Dewey.  These plus many other initiatives have lead me to conclude that we are moving to the next stage – for instance the British Library and Deutsche Nationalbibliothek are starting to cross-link their data, and the Library of Congress BIBFRAME initiative is starting to expose some of its [very linked data] thinking.

 Of course the other major initiative was that publication of Linked Data, using Schema.org, from within OCLC’s WorldCat.org, both as RDFa embedded in WorldCat detail pages, and in a download file containing the 1.2 million most highly held works.

 The need to extend the Schema.org vocabulary became clear when using it to mark up the bibliographic resources in WorldCat. The Book type defined in Schema.org, along with other types derived from CreativeWork, contain many of the properties you need to describe bibliographic resources, but is lacking in some of the more detailed ones, such as holdings count and carrier type, we wanted to represent. It was also clear that it would need more extension if we wanted to go further to define the relationships between such things as works, expressions, manifestations, and items – to talk FRBR for a moment.

The organisations behind Schema.org (Google, Bing, Yahoo, Yandex) invite proposals for extension of the vocabulary via the W3C public-vocabs mailing list.  OCLC could have taken that route directly, but at best I suggest it would have only partially served the needs of the broad spread of organisations and people who could benefit from enriched description of bibliographic resources on the web.

So that is why I formed a W3C Community Group to build a consensus on extending the Schema.org vocabulary for these types of resources.  I wanted to not only represent the needs, opinions, and experience of OCLC, but also the wider library sector of libraries, librarians, system suppliers and others.  Any generally applicable vocabulary [most importantly recognised by the major search engines] would also provide benefit for the wider bibliographic publishing, retailing, and other interested sectors.

Four months, and four conference calls (supported by OCLC – thank you), later we are a group of 55 members with a fairly active mailing list. We are making progress towards shaping up some recommendations having invested much time in discussing our objectives and the issues of describing detailed bibliographic information (often to be currently found in Marc, Onix, or other industry specific standards) in a generic web-wide vocabulary.  We are not trying to build a replacement for Marc, or turn Schema.org into a standard that you could operate a library community with.  

Applying Schema.org markup to your bibliographic data is aimed at announcing it’s presence, and the resources it describes, to the web and linking them into the web of data. I would expect to see it being applied as complementary markup to other RDF based standards such as BIBFRAME as it emerges.  Although Schema.org started with Microdata and, latterly [and increasingly] RDFa, the vocabulary is equally applicable serialised in any of the RDF formats (N-Triples, Tertle, RDF/XML, JSON) for processing and data exchange purposes.

My hope over the next few months is that we will agree and propose some extensions to schema.org (that will get accepted) especially in the areas of work/manifestation relationships, representations of identifiers other than isbn, defining content/carrier, journal articles, and a few others that may arise.  Something that has become clear in our conversations is that we also have a role as a group in providing examples of how [extended] Schema.org markup should be applied to bibliographic data.

I would characterise the stage we are at, as moving from the talking about it to doing something about it stage.  I am looking forward to the next few months with enthusiasm. 

If you want to join in, you will find us over at http://www.w3.org/community/schemabibex/ (where you will amongst other things on the Wiki find recordings and chat transcripts from the meetings so far).  If you or your group want to know more about Schema.org and it’s relevance to libraries and the broader bibliographic world, drop me a line or, if I can fit it in with my travels to conferences such as ALA, could be persuaded to stand up and talk about it.