tag:blogger.com,1999:blog-16081779.post113972651225829310..comments2023-10-28T09:24:38.420+01:00Comments on iPhylo: Rob McCool on Rethinking the Semantic WebRoderic Pagehttp://www.blogger.com/profile/00269598293846172649noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-16081779.post-1139794019848341202006-02-13T01:26:00.000+00:002006-02-13T01:26:00.000+00:00I am pretty sure I agree with you that there is un...I am pretty sure I agree with you that there is unlikely to be support for morphological characters described in RDF. <BR/><BR/>I don't even believe, unlike Kevin Thiele, the architect of LucID, that states should get GUIDS, though maybe characters should. However, my present thinking is that constrained vocabularies and descriptions are the right granularity of descriptive data on which to put GUIDs, not characters. <BR/><BR/>Ironically, when the tdwg SDD effort http://wiki.cs.umb.edu/twiki/bin/view/SDD/WebHome<BR/>began I argued for RDF and was beaten down on the grounds that it was immature and too complex (I always found this an odd argument, because it entailed that we should start all over, and that's hardly starting from a mature base...As to complexity, I quickly learned that morphological characters and states are in fact complex objects, and I am not alarmed at the complexity of what we produced in XML Schema, including some tools in support of generating import/export software and supporting developers of same).<BR/><BR/>All that said, the question of RDF for taxon and specimen descriptions, rests, for me, on the large number of arguments out there that RDF hasn't caught on much for anything other than discovery, and maybe not much for that, outside the academic world. I cited McCool because he is the most qualified I've ever seen advance the argument he does. I still suspect RDF is probably pretty good at most of what we did in SDD. However, SDD was designed as an <I>exchange</I> language for descriptions. Most holders of descriptive data will still need to hold that data in databases, which are particularly unlikely to be triple stores for a long long long time. At least one blogger claims that no RDF enthusiast has been able to show him a single encoding of <I>content</I> in RDF<BR/>http://www.zacker.org/the-battle-for-the-semantic-web-rdf-vs-xml, and I rather suspect that no scientific publications, including those with GUID metadata in RDF or any other description language, have ever been coded in such a language, at least other than as a demonstration. I'd be very interested to see what it takes to code a systematics publication in RDF. Being part of the same project you mention Donat Agosti is working on about legacy ant literature, I know what a daunting task this is if you haven't <I>started</I> from a database of descriptive data.<BR/><BR/>p.s. Please complain to your blogmaster about how the platform gratuitously truncates URLs making it pretty hard to follow links. At least for me.Anonymousnoreply@blogger.com