Hi Steve, and all RDFTM'ers

First : Good job, congratulations!

A first comment, guess what, on the section on identity :))
(1) If the topic has one or more subject identifiers and no subject locators, one subject identifier (chosen at random) becomes the URIref of the resource. Additional subject identifiers become owl:sameAs properties.
And in the final section on non-deterministic rules

(2) Topics with multiple identifiers — because there is no deterministic way to choose the identifier to use as the URIref of the resulting resource.

I would say that this is a contrario a completely deterministic situation.

TM2RDF : A topic with several subject identifiers should map to *one RDF resource per subject identifier*, those being linked to each other by owl:sameAs relationships. There is no choice to be made (random or not) of one URIref over another. Actually there is one single resource, but chosing one URIref as an arbitrary "preferred one" so to speak, does not make much sense.
Does this rule break the general principle "one TM subject = one RDF resource"? I don't think so, since resources linked by owl:sameAs properties just seem to be different because they have distinct URIref, but are actually one single resource. http://www.w3.org/TR/owl-ref/#IndividualIdentity is crystal clear about that.
I suppose you are well aware of the consequences in terms of the resulting species of OWL, of this systematic use of owl:sameAs. If the resources generated are classes or properties, they will also be individuals due to the semantics of owl:sameAs, so you are in OWL-Full. What about a possible escape rule in that case, which would be to generate owl:equivalentClass or owl:equivalentProperty predicates instead of owl:sameAs? Did you discuss this option?

RDF2TM : When two (or more) (individual) resources are linked by a owl:sameAs property, they generate a single topic bearing the URIref of those resources as subject identifiers.
Same question as above. How many topics do you generate from resources linked by owl:equivalentClass or owl:equivalentProperty?

Best

Bernard


Bernard Vatant

Knowledge Engineering

Mondeca
3, cité Nollez 75018 Paris France

Tel. +33 (0) 871 488 459 

Mail: bernard.vatant@mondeca.com

Web: www.mondeca.com

Blog : universimmedia.blogspot.com



Steve Pepper a écrit :
I am pleased to announce the availability of the first draft of the Guidelines for RDF/Topic Maps Interoperability for review by the SWBPD Working Group and the ISO Topic Maps Working Group:
 
   http://www.ontopia.net/work/guidelines.html
   http://www.ontopia.net/work/guidelines.pdf
 
The next meeting of the editors is scheduled for February 21st and we would be grateful for as much feedback as possible before then.
 
The current draft is essentially complete, except for a number of issues (all clearly marked in the document), and the section on the formal specification of the translation rules (5. Translation guidelines: formal rules). We have not yet settled on a formalism, so we would appreciate input on (1) whether we really need one (perhaps section 3. Informal Guidelines is sufficient), and (2) what formalism the WGs think might be appropriate.
 
In addition to comments on the details of the translation rules, the examples, and the general approach, we would like feedback on whether the SWBPD thinks this document should aim to become a Recommendation or just a Note. My personal opinion is that status as a Recommendation would do a lot to enhance the "prestige" of the Guidelines and thus encourage wider adoption.
 
I would like to draw the attention of members of the OEM Task Force to section 3.6.2 N-ary relationships in particular. As you will see, we have based our approach on the work done by Natasha, Alan and Pat in the document Defining N-ary Relations on the Semantic Web (latest draft at http://smi-web.stanford.edu/people/noy/nAryRelations/n-aryRelations-2nd-WD.html). It seems to us that we only need to define a single class (which we have called rdftm:N-aryProperty, for consistency with the rest of the RDFTM Guidelines) in order to both represent Pattern 1 (A and B) and provide the guidance necessary to achieve RDFTM interoperability. We would appreciate your feedback on this.
 
The work of the editors has been taking place using the University of Bologna Wiki at http://tesi.fabio.web.cs.unibo.it/cgi-bin/twiki/bin/view/RDFTM. Minutes of our conference calls are available at http://tesi.fabio.web.cs.unibo.it/RDFTM/MinutesOfConferenceCalls.
 
We look forward to receiving your comments.
 
Finally, let me take this opportunity to apologize for my lack of active participation in the SWBPD WG during the last months: I have been off sick for quite a while. I will try to ensure that at least one of the RDFTM editors participates in WG telecons from now on.
 
Best regards,
 
Steve

--
Steve Pepper <pepper@ontopia.net>
Chief Strategy Officer, Ontopia
Convenor, ISO/IEC JTC 1/SC 34/WG 3
Editor, XTM (XML Topic Maps 1.0)