After a summer break I’m back into blogging. Before continuing my tutorial, I’d like to talk about the Dublin Core conference, which took place last week in Lisbon and was collocated with iPres. Both conferences had a heavy presence of people from the digital libraries, who provided different perspectives on many Semantic Web topics. Also, when registered in one event, people could freely sneak in the talks of the other and vice versa, providing a flexible schedule for all the attendants.
My main reason to be there was a 3 hour tutorial about PROV and the mapping to Dublin Core the first day. 22 people attended and participated a lot during the session, which made it more interesting. There was some discussion and I received some questions, mainly from librarians, about PROV. The mapping received some critics but many participants considered it valuable as an entry point to use PROV in their datasets. Also, I learnt that now there is an effort to align PREMIS (http://www.loc.gov/standards/premis/) to PROV. For those interested, the slides of the tutorial can be accessed here: http://www.slideshare.net/dgarijo/provo-tutorial-dc2013-conference.
My highlights of the conference (there were parallel sessions, see the full program here):
Gildas Illien’s keynote (Darling, we need to talk), where he emphasized the different views of the preservationist vs bibliographists, and summarized the process of making the BNF (French National Library) compliant to the Linked Data standards. As always, an important conclusion is to keep the complexity of the system (e.g., RDF) hidden from the users.
The vocabulary preservation session, where the LOCKSS system (Lots of copies keep stuff safe) was briefly presented and discussed, along with other projects like Memento, which aim to provide entities at a given point in time. The people from LOV were also present. Unfortunately the discussion ended without a clear conclusion, so Ivan Herman proposed to create a new W3C community group where the discussion continues with all the parties interested.
The poor versus rich session on vocabularies used for exposing Linked Data of digital libraries across different countries like Spain, France and Germany. All the participants of the panel introduced how their systems had adopted Linked Data, although in the end there was not much discussion about the vocabularies used.
The last keynote by Carlos Morais Pires about the Horizon 2020 call and how it will be related to e-Infrastructure. Data will have to become an infrastructure itself, by sharing and federating it. “Homeless data becomes no data after all”. This is very interesting since many proposals in this regard (related also to Research Objects) will arise during the next year.
Finally, one of the most interesting sessions in my opinion was the one about Schema.org. Dan Brickley stated how Schema.org does not aim to be a top level ontology, but an empirical vocabulary to communicate with search engines and improve the visibility of your web pages (Google accepts annotations in microdata and RDF-a). The vocabulary is still incomplete in many aspects, but different organizations and community groups are starting to propose terms for adoption (the OCLC people had recently proposed one to track the bibliography items). However this task is not easy, as the need for a term has to be motivated by demonstrating that any of the existing ones don’t fit its purpose.
And I’d like to end this post with two images: the Fado concert given during the dinner plus the reception by the Tuna of the university. Special thanks to Stuart Sutton and José Borbinha, the organizers.