Archive for December, 2008|Monthly archive page

Microformats in RDF

I’ve been  to VoCamp Galway 2008 last week, which has been my first VoCamp. I have to say it has been a fantastic experience. It seems like the VoCamp formula just works out of the box: you let in a bunch of vocabulary geeks inside a closed room and they immediately get down to fixing some of the key outstanding questions surrounding vocabularies on the Semantic Web (and in their spare time, they create some new vocabs).

The primary issue I set out to discuss was the state of RDF vocabularies for microformats. Why is this important? A growing number of tools exist that convert microformats into RDF but in case everyone chooses different representations it will be significantly more difficult to interoperate. So in a small group we set out to document what we believe are the best practice mappings from microformats to RDF.

We started by collecting the existing RDF vocabularies for stable microformats and documented them on the SemanticWeb.org wiki. Alone this has been an incredibly useful exercise since we uncovered some pretty immediate targets for fixing:

  • hCard. The fact there are two VCard ontologies lingering on W3C’s website [1,2]. It’s clear to the trained eye that one is deprecating the other, but unfortunately it doesn’t say so anywhere… I’ve written an email about this to the friendly folks at W3C.
  • hCalendar…same here. it seems the RDF Calendar group couldn’t agree at some point on the design, forking the ontology into two… We now took this discussion to the RDF Calendar mailing list, with the intention to consolidate things.
  • rel-license… we have three terms for saying ‘this is the license for this doc’: xhtml:license (which is the default namespace in RDFa), dc:license and cc:license.
  • hAtom… Again, there seem to be multiple chooses…we only got as far to list the ones we knew about.
  • hResume… In this case we haven’t found a direct mapping: there are various resume ontologies but they all do more or something slightly different than hResume.

This is about as far as we got, but I’ve seen that in the past week already people have added some pretty useful info to the Wiki at semanticweb.org. If you have something to contribute as well, please do so!