RE: Review of the "SKOS Reference" Last Call Working Draft [ISSUE-138]

Dear Sean, Dear WG!

This is a collective answer to your answers concerning issues ISSUE-136,
ISSUE-137 and ISSUE-138. The reason is that all three issues and your
respective answers are pretty similar in their form.

Let me say the following, just in order if there was a misunderstanding
regarding the targets behind my comments. I did /not/ want the working group
to refer to OWL 2, or to change the given semantic conditions. So postponing
those issues is a solution, but I would also be satisfied if these issues
are just closed.

However, I believe that in all these cases, where a semantic condition
cannot be related to RDF(S) semantics or OWL, it should be stated
explicitly, at least /that/ it cannot be done. I can see that Section 1.7.1.
makes a generic comment in this direction. But having one additional, and
perhaps a bit more concrete sentence at the respective semantic conditions
each would be helpful. Comments such as "not possible in OWL" or "can only
approximated to some degree", or similar comments. In addition to helping
"ordinary" readers to understand what's up with a particular semantic
condition, it would also be a good hint for future working groups. 

But I leave the decision to add such additional information or not to the
working group.


>-----Original Message-----
>From: Sean Bechhofer []
>Sent: Friday, October 17, 2008 5:53 PM
>To: Michael Schneider
>Cc: SWD Working SWD
>Subject: Re: Review of the "SKOS Reference" Last Call Working Draft
>Dear Michael
>Thank you for your comments [1,ISSUE-138]:
>Such a kind of functional property characteristics distinguished by
>tags cannot be expressed in OWL Full. I suggest to say this somewhere
>in the
>Again, I want to hint you to rdf:text, in combination with OWL 2,
>which is going
>to allow for specifying data ranges consisting of all plain literals
>having a
>given language tag. A somewhat more general feature was the
>resolution of
>issue-71 of the OWL WG, accepting this proposal:
>0053.html>. But,
>AFAICS, one would need to build a <=1-qualified cardinality
>restriction (new
>feature in OWL 2) on each such language-data range in order to
>simulate a
>functional property of the kind above. So this is probably all a bit
>As you point out, there are some constraints in the SKOS data model
>that we are unable to express in OWL (some of these /may/ be
>addressed by OWL 2, but in the current SKOS specification we are
>avoiding reference to work in progress). In such cases, the
>constraints are expressed in prose in the document.
>Statements to this effect are made in Section 1.7.1 of the LC draft.
>Do you feel these are sufficient, or do we need to further elaborate
>this point?
>The Working Group propose to *postpone* this issue, indicating that
>this may be an area that future groups may wish to return to. Are you
>willing to live with this?
>	Sean Bechhofer
>	Alistair Miles
>Sean Bechhofer
>School of Computer Science
>University of Manchester

Dipl.-Inform. Michael Schneider
FZI Forschungszentrum Informatik Karlsruhe
Abtl. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Web  :

FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts
Az: 14-0563.1 Regierungspräsidium Karlsruhe
Vorstand: Rüdiger Dillmann, Michael Flor, Jivka Ovtcharova, Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus

Received on Wednesday, 22 October 2008 21:06:00 UTC