finally a bnode with a uri

SeenOn - Timestamp or State of Mind?

fun stuff from #microformats, comments on e/RDF/a wrt to Microformats
<tommorris> Every time I see a movie from now on,
  I'm adding the IMDB URL to my FOAF file.
<briansuda> with what predicate?
<briansuda> seenOn, is that a timestamp or a state-of-mind?
(microformats(!) irc channel)

Now, who said RDF was less real-word-ish than microformats?

Related link (wrt to movies, not toxics): Microformats 80%, RDF 20% by Tom Morris about the longtail utility of (e)RDF(a). Wanted to state something like this for some time. After implementing a Microcontent parser (part of the next ARC release) that creates a merged triple set from eRDF and Microformats, I can't say anymore that MFs don't scale (even though making the meaning of nested formats explicit is sometimes tricky). I was really impressed by the amount of practical use cases covered by them (Listings and qualified review ratings even go beyond the demos I've seen in RDFer circles). However, there is still a lot of room for custom RDF extensions that can be used to extend microformatted HTML. Skill levels are just one of many longtail examples: They are currently not covered by hResume, but available in Uldis' CV vocab.

The important thing IMO is that RDFers should not forget to acknowledge the amazing deployment work of the MF community and focus on what they can add to the table (storage, querying, and mixing, as a start) instead of marketing RDF-in-HTML as an alternative, replacement, or otherwise "superior" (likewise the other way round, btw.). I think we also shouldn't overcharge the big content re-publishers. When maintainers of sites like LinkedIn or Eventful get bombed with requests to add different semantic serializations to their pages, they may hesitate to support any of them at all. For most of these mainstream sites, Microformats do the job just fine, and often better. Why should people for example have to specify namespaces when a simple, agreed-on rel-license does the trick already? (We could still use RDF to specify the license details, and even the license link is only a simple conversion away from RDF.)

Comments and Trackbacks

whats this about specifying namespaces? i had to use DOAP (including some custom extensions), FOAF, SIOC, DC, and DOAC to even get a basic resume going. absolute URIs beat some vague notions non-deferencable quasi-agree-upon basket of rel-tag overload. imo..
Comment by c on 2007-02-11 03:00:19 UTC
c, maybe I wasn't clear enough about what I tried to state, sorry in that case. I was referring to those use cases where you don't need custom RDF vocabularies (e.g. for license links). And even for resumes you can get very far with MFs.

Note that I'm not talking about "namespaces vs. MFs". Of course I want URI-identified terms in my app, but if there is a practical mapping from HTML to RDF, I don't necessarily have to have them in my HTML. The parser I wrote ends up with just the same mix of vocabularies you'd have if you had created an RDF description in the first place.

Things like rel-tag overload will only happen when we start to redundantly squeeze MFs, eRDF, and RDFa markup into our Web documents where one of them would have been enough.
Comment by Benjamin Nowack on 2007-02-11 11:53:16 UTC
0 comments are currently in the approval queue.

Comments are disabled for this post.

Earlier Posts


No Posts found