- From: Dan Brickley <danbri@w3.org>
- Date: Mon, 9 Apr 2001 16:47:35 -0400 (EDT)
- To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- cc: <danny@panlanka.net>, <drew.mcdermott@yale.edu>, <www-rdf-logic@w3.org>
On Mon, 9 Apr 2001, Peter F. Patel-Schneider wrote: > It is very fashionable to put down lawers (and, in this group, it also > appears to be fasionable to put down logicians). However, legal training > and experience is necessary to craft documents and agreements that will > 1/ stand up in court, and 2/ not cause more problems than they solve. I > view the situation with respect to representation as quite similar: It > takes logical training and experience to craft representations and > interfaces that will 1/ actually say and do what you want them to (when not > helped along by human-level understanding), and 2/ not cause more problems > later on than they solve now. > > Perhaps those of us who are arguing for logical sophistication have not > been as explicit as we might in making our case. However, Drew, Pat, and I > have seen firsthand the problems that can arise when logical precepts are > ignored. (I have even seen such problems with respect to my own work.) > > All this is not to say that it is always necessary to take care of all the > logical precepts in designing a representation system. There are many > useful representation systems that are on shakey ground. However, such > systems run into problems when they are used as components of other > systems, or are used by people (or systems) beyond the initial core group, > or are used in ways that were not envisioned or completely thought through > by the initial core group. If you think that a web representation system > that is supposed to form the basis of all semantic web work doesn't fit > into any of these categories then you don't need a firmly-grounded system. > If, on the other hand, [you can fill in this part] .... > > Peter F. Patel-Schneider I absolutely agree. We created www-rdf-logic as a forum for those who _do_ see the need to use logic when designing representation systems for the Web. This list should not be a place where logicians have to justify their work, or its relevance to W3C's Semantic Web effort. Quite how we go about organising an "after the party" cleanup of the Web (eg. URIs, the murkier corners of the RDF specs...) is another matter. I expect to see, for example, the new RDFCore WG[1] take on board some of the clarifications to RDF (eg. RDFS semantics) proposed by DAML. But I have no instinct as to how far (or how fast) we can go in formalising some of the other key components of the Web (eg. the naming model implicit in http://www.ietf.org/rfc/rfc2396.txt). Dan [1] http://www.w3.org/2001/sw/#RDF-Core-WG
Received on Monday, 9 April 2001 16:48:02 UTC