- From: Benjamin Nowack <bnowack@semsol.com>
- Date: Thu, 7 May 2009 18:20:15 +0200
- To: Richard Cyganiak <richard@cyganiak.de>
- Cc: semantic-web at W3C <semantic-web@w3c.org>
[did some CC cleanup, hope that's ok] On 07.05.2009 14:10:34, Richard Cyganiak wrote: >4. Yahoo reports 1B+ hcard microformats on the Web. All microformat-to- >RDF converters I'm aware of map them to the new namespace (e.g. any23, >Swignition, SearchMonkey). All ARC-based ones don't (not sure how many there are, though ;). >I conclude that the old namespace is obsolete and should be ignored >and avoided, in the interest of converging on a single namespace. Individual conclusions probably don't have much effect here (unless you're an html5 spec editor ;) One reason why I preferred the old schema, for example, was the ability to attach selected terms to foaf:Persons. And the 2006 effort felt rather "discontinued after lack of agreement" than "done so far" to me. IIRC the main issues were 1) vcard disjoint from person?, 2) ordered values (seq vs list), 3) fon/fax/email types, no exact 1-to-1 mapping possible? (value annotations vs. structured values vs. interpretation props) There were good arguments pro and contra each solution. In the meantime, as Danbri mentioned already, OpenSocial and PortableContacts entered the scene. I personally started using an RDF version of OpenSocial (e.g. in Knowee) and find it quite handy. OpenSocial comes with PoCo fields which in turn contain a clean vcard term set. If we could talk the Osoc folks into serving an RDF Schema from their site, we could close many issues around vcard and names in foaf. But again, this is just one person's opinion and won't have much effect ;) If I had anything to conclude, then I'd say that we could benefit from some use-case-centric education site where you can pick a use case (not a vocab!) and get examples illustrating the combination of possible RDF schemas to encode the use case ("simple address book", "people with multiple addresses", "social network contacts", ...). Each example could mention alternative terms or approaches, and maybe SPARQLy mappings between the different vocabs. It would help us, is probably nice for beginners, and it would also show that the existence of overlapping vocabs with slightly different focus or interpretation of a source domain/format doesn't mean "fundamentally flawed approach" which we often hear from those centralization lemmings. Cheers, Benji -- Benjamin Nowack http://bnode.org/ http://semsol.com/ <spam> RDF/PHP talk & workshop in NYC: http://bit.ly/VfLhB </spam>
Received on Thursday, 7 May 2009 16:20:53 UTC