- From: Richard Newman <rnewman@twinql.com>
- Date: Thu, 28 Feb 2008 19:55:52 -0800
- To: Paul Tyson <phtyson@sbcglobal.net>
- Cc: SW-forum list <semantic-web@w3.org>
> From the "metadata is good, ontology is bad" camp, see: > > http://www.shirky.com/writings/semantic_syllogism.html Beware of that much-hyped essay; it's incomplete and full of straw men, which tends to weaken its point. I trust Paul Ford and Peter Van Dijck far more: http://www.ftrain.com/ContraShirky.html http://poorbuthappy.com/ease/semantic/ and there are other analyses: http://www.intertwingly.net/blog/1641.html > Azamat's suggestion should be taken seriously by everyone > interested in the semantic web. Given that he advocates a global ontology (depending on how you define the term, this could be anything from "ridiculous" through to "possible impractical or incorrect" to "probably necessary"), and seems to heavily favor deeply abstracted methods of knowledge representation, I think his suggestion should be taken with a pinch of salt. (As should mine -- skepticism is a valuable trait.) More knowledge is almost always a good thing -- I agree with you there; people should go and read those links! -- but the agenda behind his suggestion (which pitches us towards an abstracted, formal, semiotics-based knowledge representation system, backed by a gigantic, "correct", global ontology) is quite likely impractical, and possibly detrimental to the state of the world. I'd rather use imperfect tools, and improve them over time, than hunt with flint spears until the perfect knowledge representation system comes along. To return to the original email: SKOS is doubtless imperfect, but it's better than the current alternative (rolling my own), and it doesn't prevent Azamat from going off and building something better. -R
Received on Friday, 29 February 2008 03:56:03 UTC