RE: Characterising OWL consistency checkers

Some points:
- I like Ian's words for consistency checker, and will be happy to accept
them as editorial improvements. If I understand correctly they depend on
some changes in the S&AS; we currently have consistent defined in S&AS
rather than test cases, and test cases points to S&AS. This was a WG
decision (at January f2f), so I would prefer a WG decision to change the
text to give an in-line definition of "consistent".

- Ian you give DECIDABLE and UNDECIDABLE as if there is some markup, e.g. a
link to a glossary? Did you have something in mind?

- If we have a non-empty list of supported datatypes I would like
rdf:XMLLiteral in it.

- Ian suggests integer as a possible required datatype. I have no problem
with this but wish to clarify that this is an infinite set of integers.

- if we support all derived types of supported types as supported then we
have a problemette with xsd:ID and xsd:IDREF that we do not recommend, but
are derived from xsd:string.

- if we don't support all derived types of supported types, then I suggest
we should support all finite derived types of xsd:integer.

Jeremy



> -----Original Message-----
> From: www-webont-wg-request@w3.org
> [mailto:www-webont-wg-request@w3.org]On Behalf Of Ian Horrocks
> Sent: 08 March 2003 20:12
> To: Jim Hendler
> Cc: www-webont-wg@w3.org
> Subject: Re: Characterising OWL consistency checkers
>
>
>
> On March 7, Jim Hendler writes:
> >
> > At 18:02 +0000 3/7/03, Ian Horrocks wrote:
> > >On March 7, Jim Hendler writes:
> >
> > >[snip
> >
> >
> > >Because ALL datatypes/values are interpreted, an OWL reasoner MUST
> > >"deal with" all datatypes. This isn't a big overhead because
> > >unsupported datatypes can be treated in a uniform and very lightweight
> > >manner. If a reasoner claims to support a given datatype, AND it
> > >claims to be complete, then it has to do more work w.r.t. that
> > >datatype (the MT formalises just what it has to do).
> >
> > yes, but remember that many of our readers won't know this, and thus
> > I just want to make sure someone says it in novice-understandable
> > form in our document.  I did very much like what you wrote, wasn't
> > being critical
>
> OK - fair enough. I can try to come up with a few more words.
>
> >
> > >
> > >>    but we need some decision on this
> > >
> > >The decision as to which datatypes (if any) MUST be supported by all
> > >OWL reasoners is obviously one that has to be made by the working
> > >group. My suggestion is that we make this a fairly short (possibly
> > >zero length) list.
> >
> > actually, I think your ACTION item was to take a first stab at making
> > a suggestion.  If you want to suggest zero length, that woudl address
> > your action.  I probably prefer string and integer, but could live
> > with none.
>
> Agreed. Wording could be:
>
>   An OWL consistency checker MUST support at least the following XMLS
>   datatypes: integer, string.
>
> There is an issue as to whether we want to include all the sub-types
> if integer in this list, e.g., int and byte. There is also the
> question of whether we want to add anything about additional datatypes
> that OWL "should" support? E.g.:
>
>   An OWL consistency checker SHOULD also support the following XMLS
>   datatypes: ...
>
> Ian
>
>
> >
> > Guus, can you make sure this goes on WG agenda to get resolved on 13th?
> >   cheers
> >   JH
> >
> > --
> > Professor James Hendler				  hendler@cs.umd.edu
> > Director, Semantic Web and Agent Technologies	  301-405-2696
> > Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
> > Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
> > http://www.cs.umd.edu/users/hendler
>
>

Received on Monday, 10 March 2003 08:52:30 UTC