I must admit I was a little skeptical of the timing when I accepted the invitation to provide the keynote for a Linked Data session – on the last day of IFLA 2010 – at 8:30 in the morning – in August – on a Sunday. Who was going to want to get up at that time, on the day they were probably going to leave beautiful Gothenburg, to hear me witter on about the Semantic Web and the obvious benefits of Linked Data for libraries? A few minutes before the start, I was beginning to think my skepticism was well founded, viewing the acres of empty seats laid out in their menacing ranks in front of me. But then almost as if from nowhere, the room rapidly filled and by the time I took the stage we had something approaching a full house. As you can see from my iPhone snap below, we ended up with a significant group (I lost count at about 250) of interested librarians.
So was it worth them turning up at such an unsociable time? I obviously can’t speak for my session, but I believe it was well worth turning up. We had a series talks which varied from the in-depth technical/ontological spectrum to the rousing plea to open up your data now – and don’t hamper it with too much licensing.
First on after my session was Gordon Dunsire from the University of Strathclyde who gave us some in depth reasoning as to why we needed
complex detailed ontologies based upon standards like RDA, FRBR, FRAD, and RDA to describe library resources in RDF for the Semantic Web. To represent the full detail that catalogers have, and want to, provide for resource description I agree with him. I also believe that we need to temper that detailed view by including more generic ontologies in addition. People from outside of the library world, dipping into library data [with more ways to describe a title than there are flavors of ice cream], will back off and not link to it unless the can find a nice friendly dc:title or foaf:name that they understand.
Some of the other speakers that I caught included Patrick Danowski’s entertaining presentation entitled “Step 1: Blow up the silo!”. He took us through the possible licenses to use for sharing data, only to conclude that the best approach was totally open public domain. He then went on to recommend CC0 and/or PDDL as the best way to indicate that your data is open for anyone to do anything with.
Jan Hanneman from the German National Library delivered an interesting description [pdf]of the way they have been publishing their authority data as Linked Data, and the challenges they met on the way. These included legal and licensing issues, around what and under what terms they could publish. Scalability of their service, being another key issue once they move beyond authority data.
All in all it was an excellent Sunday morning in Gothenburg. I presume the organizers of IFLA 2011 will take note of the interest and build a larger, more convenient, slot in the programme for Linked Data.