- From: Je'ro^me Euzenat <Jerome.Euzenat@inrialpes.fr>
- Date: Mon, 23 Oct 2000 09:02:25 +0200
- To: "Hart, Lewis" <lhart@grci.com>, Pierre-Antoine CHAMPIN <pachampi@caramail.com>, Jeff Heflin <heflin@cs.umd.edu>
- Cc: www-rdf-logic@w3.org, "Emery, Pat" <pemery@grci.com>
Hello, I must first say that I subscribe to Deborah L. McGuinness tentative ontology of closing issues. Moreover, my initial though when initiating the threads were not about belief and trust, but more oriented toward software (ontology) engineering. I.e. how can I be sure that the weakest part of the reused classes will not damage the strongest. In his message (RE: Re[1] DAML-ONT: the case for closedness) of 20/10/00, Hart, Lewis wrote: >and also.. >Jeff Heflin said: > > >Here's what it looks like in RDF (assuming http:/www.daml.org/damlont > >and http:/www.commerce.org/econt are the daml and ec namespaces from > >above): > > > > <rdf:Description about="http://www.daml.org/damlont#Thing"> > > <rdfs:subClassOf resource="http://www.commerce.org/econt#Computer" >/> > > </rdf:Description> > > > >This is an unfortunate consequence of RDF's "anyone can say anything > >about anything" property." Thus, trust becomes 100 times more important > >for ontologies than it does for ordinary RDF documents. If you > >mistakenly trust a plain RDF document, at worst you may believe a > >hundred false assertions. If you mistakenly trust an ontology, all of > >your subsequent beliefs could turn out to be wrong! >The "bad:" ontology can state whatever it likes. The question is >do you believe it. The situation I had in mind at the beginning is not that simple. The point is in having three Ontologies: - myO, your ontology for whatever that you will build, - bigO, the well-known ontology that anybody uses in your field and that you know you can trust very much. - theotherO, the ontology that is provided by the little startup you don't know if you must trust but that provides you with the lastthing widgets in ebusiness and that claim to be bigO compliant. Well, you would like to be able to use these in such a way that even if theotherO is flawed in some parts (that maybe you do not use), this cannot affect the well-established concepts that you know from bigO. Unfortunately, due to the open character of the definitions, someone can come and restrict a definition of a concept in bigO (and this can be silent because it does not raise incoherence, just restrict one or two extensions). So the point is: I trust bigO for sure. I am not really confident about theotherO, but their new concepts are interesting and they are bigO compliant (let say that I can still use the concepts I know very well from bigO with their stuff). Now, DAML does not provide any way for the bigO makers to force some compliance. The guys from theotherO.com are not BAD guys as have been said: they just believe that Thing=Nothing and thus everything is a computer. And, after all, this is a model of the bigO ontology! This can be less dramatic and less obvious (as soon as you begin to consider an object, you will realize that Thing=Nothing) but more insidious. So I really would like to see a way to have some restrictions of that kind in some DAML definition. >The good things about this are: > >1) Anybody can still say anything. >2) There is no central authority. > 2a) Authorities will arise out of usage. >3) The scope of bad a ontology is limited. > 3a) Search engines can cut them off at the root if desired. >4) Information can still be found. The last point can be debated, since a bad ontology can describe very relevant information (but maybe you mean that the sites will be annotated with several differing ontologies?). >The Bad things are: > >1) Systems will have to delineate between what it knows about > and what it believes. >2) Keeping track of this could be resource intense. >3) Some sites will think everything is a computer. And? Isn't it correct? ;-) -- Jérôme Euzenat __ / /\ INRIA Rhône-Alpes, _/ _ _ _ _ _ /_) | ` / ) | \ \ /_) 655, avenue de l'Europe, (___/___(_/_/ / /_(_________________ Montbonnot St Martin, / http://www.inrialpes.fr/exmo 38334 Saint-Ismier cedex, / Jerome.Euzenat@inrialpes.fr France____________________/ Jerome.Euzenat@free.fr
Received on Monday, 23 October 2000 03:03:29 UTC