- From: Dan Brickley <danbri@w3.org>
- Date: Fri, 4 Oct 2002 08:13:33 -0400 (EDT)
- To: Bill de hÓra <dehora@eircom.net>
- cc: "'Paskin, Norman (DOI-ELS)'" <n.paskin@DOI.org>, <www-rdf-interest@w3.org>
On Fri, 4 Oct 2002, Bill de hÓra wrote: > > If RDF is going to be useful in dealing with real world > > problems then it will be necessary to deal with complex > > descriptions that require more expressive data models able to > > differentiate between agents, documents, contexts, events, > > and the like. > > Hmm, what do you mean exactly by "real world"? > > You're assuming that people have the time, expertise, consistency, and > funds to markup their information with ontologies. Unless you have a > very simple way of allowing people to do that with incurring upfront > costs, you run right into the active user paradox, one that seems to > plague RDF's adoption. > > > > Rather than DC, ontology-based metadata systems > > that are based on structured data models: to name a few, > > MPEG-21's RDD; SMPTE; CIDOC's CRM; the library world FRBR, > > etc; and tools which provide a means of mapping these like > > indecs, the ABC model etc. > > You're generalizing again. I'm not sure what this has to do with RSS... > ? FWIW I've worked on RSS 1.0, DC, ABC as well as tracking the INDECS, CRM etc work. There's no conflict: RDF was designed so that all such vocabs could be deployed within the same overarching framework. And RSS 1.0 was designed so that wide range of descriptive vocabulaies could be deployed within RSS feeds. Oftentimes, Dublin Core is plenty complex enough. Sometimes, a more explicitly modelled representation is worth the extra effort it imposes on consumers and producers of the feeds. Dan
Received on Friday, 4 October 2002 08:13:36 UTC