- From: Didier VILLEVALOIS <dvillevalois@techmetrix.net>
- Date: Mon, 22 Apr 2002 18:25:47 +0200
- To: Jeff Heflin <heflin@cse.lehigh.edu>, Drew McDermott <drew.mcdermott@yale.edu>
- Cc: drager@bbn.com, www-rdf-logic@w3.org
Hello, Shouldn't RDF support something like daml:imports in its core ?? daml:imports would be a subPropertyOf 'rdf:imports'. As rdf enables to speak about the same things in different places, it needs a such mechanism. This would enable everybody to make rdf documents that are perspectives on the semantic web. I think this should not be as high-level as being part of DAML, because everybody's semweb engine may not have it. Didier. > -----Message d'origine----- > De: Jeff Heflin [mailto:heflin@cse.lehigh.edu] > Date: lundi 22 avril 2002 18:11 > À: Drew McDermott > Cc: drager@bbn.com; www-rdf-logic@w3.org > Objet: Re: DAML Level of Effort for FY03-FY05 > > > Drew, > > Since I was the initial proponent of daml:imports on the Joint > Committee, let me address this issue. You are absolutely correct that > the imports statement must be used. Simply refering to a > namespace does > not include any ontology definitions. You must make the imports > statement explicit. Period. > The correct way to do this for instance data is: > > <rdf:Description about=""> > <daml:imports > rdf:resource="http://www.daml.org/experiment/ontology/beta/ele > ments-ont#"/> > </rdf:Description> > > The problem with using RDF namespaces to decide which schemas are > relevant is that multiple files may contain different > definitions about > the same URI. See the attached GIF for example. The URI for > Dolphin has > additional definitions in two schemas, good-schema and > bad-schema. These > definitions are simply rdfs:subClassOf statements that happen to have > orig-schema#Dolphin as their subject. The problem with simply using > namespaces is I can't say that my-doc agrees with the definitions of > Dolphin found in good-schema but not those found in > bad-schema. This is > why daml:imports was an essential component of the language. Note that > this issue is also addressed in the Web Ontology Requirements document > [1]. > > Now this does mean that there will be some apparent redundancy between > the namespace declarations and the imports declarations in any DAML > documents. This is unfortunate, but is the result of XML namespaces > being defined solely to solve the namespace problem for > languages before > the W3C was concerned with ontologies. The ideal solution would be to > have the imports introduce namespaces for each ontology (as > is done with > most pre-Web ontology languages), but alas this solution would go > against the grain of pre-existing web standards. > > Theoretically we could say that the use of a namespace > implies that you > import that ontology, but this does have the issue (raised > later in this > thread) of distinguishing ontologies from other XML > namespaces. Also, it > would not allow someone to provide completely alternate > definitions for > a common set of terms (whether this is a good thing or a bad > thing, I'm > not sure). Regardless, there is nothing in the DAML+OIL specs > that says > that such default import statements exist. Although some tools may > decide to implement this default, it is nonstandard, and is akin to to > having HTML tools that are very flexible about guessing what you mean > when you provide malformed HTML. > > Jeff Heflin > > [1] > http://www.w3.org/TR/2002/WD-webont-req-20020307/#goal-shared- ontologies Drew McDermott wrote: > > Hello, Dave -- > > In your capacity as keeper of the DARPA/SONAT ontologies, I have a > question about the "instance data." I am cc-ing www-rdf-logic because > there seems to be a general issue here about how DAML datasets work. > > These ontologies are found at > > http://www.daml.org/experiment/ontology/ > > and I don't think they're sensitive (or they would be https: instead > of http:). > > The file af-a.daml is RDF, rather than DAML, and appears to consist of > "instance data," not ontology. > > Here is the beginning of that file: > > <rdf:RDF > xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#' > xmlns:rdfs='http://www.w3.org/2000/01/rdf-schema#' > xmlns:NS0='http://www.daml.org/experiment/ontology/beta/elements-ont#' > > > > <rdf:Description rdf:about='http://www.daml.org/2001/12/factbook/af.daml#A1202'> > <NS0:country rdf:resource='http://www.daml.org/2001/09/countries/fips#AF'/> > <rdf:type rdf:resource='http://www.daml.org/2001/12/factbook/factbook-ont#Port'/> > <rdfs:label>Kheyrabad</rdfs:label> > </rdf:Description> > > My problem is this: > > Wouldn't it be much more informative if this line > > xmlns:NS0='http://www.daml.org/experiment/ontology/beta/elements-ont#' > > were supplemented by something like this: > > <daml:imports rdf:resource="http://www.daml.org/experiment/ontology/beta/elements-ont#"/> > > But it's not clear where this would go. > > Is it supposed to be the case that just mentioning an ontology's > namespace means that the file uses that ontology? I find that rather > disconcerting. For the time being, we'll make that assumption, but > this seems like a bug (or gap) in the DAML spec. There ought to be an > official way to say "This RDF file is based on ontologies X, Y, and > Z." > > -- Drew McDermott
Received on Monday, 22 April 2002 12:26:24 UTC