- From: m batsis <mbatsis@netsmart.gr>
- Date: Sat, 21 Sep 2002 14:55:43 -0700
- To: SALVETTI FRANCO <franco.salvetti@colorado.edu>
- Cc: www-rdf-interest@w3.org
On Friday 20 September 2002 18:26, SALVETTI FRANCO wrote: [...] The rande/domanin are not > constraint anymore, but just way to declare the type of a variable. I heard > that it could be changed in the future by the W3C. > > More ore less with this semantic you can say everything in RDF. No > constarints! Well, it's all about the possible inferences. The term constraint in RDF(S) is simply confusing as DanBri noted. > Is this a good way to produce an object so sophisticated as the Semantic > Web? Yes. There is a number of parts missing; see below. <lots of "do you think" questions> > Are we thinking about a small private application or at the whole WEB? Great point. Again, see below. > sorry for the long letter, but sometimes I think that we have to see the > problem from 10,000ft high (as was said by Lee)... but now seems that we > are more concerd in small problems and we have lost the general picture of > the real problem. Do you think that with RDF we can beat google and its > very siple algorithm of ranking? Well, that's why I'm on this list. RDF may not be usable for first order logic, but offers great power and the potential of widespread use. Here's some things that currently spoils the big picture of the SW: * Inability of RDF(S) to go beyond closed worlds without the use of predefined and agreed_upon ontologies. RDF(S) provides no way of assosiating an RDF instance document with a set of RDF Schemas, so an RDF processor cannot simply retreive that when it processes unknown terms. So, globality stopes right there, however, RDF authority simply has not utilized URIs to provide that (the only place where this can actually work IMHO). The problem is the neccessary mapping between URIs as objects and namespace URIs, which simply wont happen. * RDF(S) has a bad reputation of being difficult to use and deploy. This, of course, has some truth; missing parts of the picture force developers to provide hacks for some functionality that should be build-in. Other than that, you can easily write RDF; it's natural to think in it. * Vendor support is extremely low. They will not do the dirty work and they simply wont help the community gain so much information power; what will they then do with all the legacy (when compared to SW) software, including app servers and trash editors? * We are still waiting for datatyping to be resolved. * IMHO, the biggest current problem of RDF is the interpretation of URIs. They are anticipated as just unique names, but nothing besides that. URIs are actually a sort of a structured information, or rather, a description of a path in such a structure. RDF does not see it that way I think. So, RDF is currently being reduced to DCMI style techniques. I hope WebOnt will dare to push it beyond that, by applying a new semantic layer on top of the RDF spec. Manos
Received on Saturday, 21 September 2002 08:02:27 UTC