untyped URIs was RE: Agenda/Logistics, Oct 2 telecon

The outstanding issue from Oct 2 is B below

I discussed this with Peter in Florida and we have a joint proposal on B.2
and I would be happy to abstain on a resolution for B.1 following Peter's
preference.
(see below)

>
> Jim
> here is a list of test related items that I am aware of, maybe
> for under 3.3
> in the agenda
>
> [[
> 3.3 - Peter and Jeremy have noted some discrepancies between Test and
> S&AS, since both are normative, these need resolution.  There seems
> to be little dispute, but WG should approve any of these needing a
> decision
> ]]
>
>
> B) uri references without a type
> B.1) as object of annotation properties
> B.2) as user defined datatype
>
> B) uri references without a type
>
> B.1) as object of annotation properties
> Test case:
> http://www.w3.org/2002/03owlt/editors-draft/draft/proposedByFuncti
> on#AnnotationProperty-003
>
> jeremy:
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Sep/0313
> peter:
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Sep/0319
>
> My preference is:
> PROPOSE modify S&AS to ensure that urireference objects of annotation
> properties are one of
> datatypeID, classID, individualID, ontologyID, datavaluedPropertyID,
> individualvaluedPropertyID, annotationPropertyID, ontologyPropertyID.
>
> B.2) as user defined datatype
> Test case:
> http://www.w3.org/2002/03owlt/editors-draft/draft/proposedByIssue#I5.8-016
>
> (also note test I5.8-013, I5.8-014, I5.8-015 which are related)
>
> jeremy:
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Sep/0154
> peter:
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Sep/0181
>
> I would be happy with either:
> - modifying S&AS to require an
>   ddd rdf:type rdfs:Datatype .
>  triple
> or
> -modifying S&AS to be underdefined in this area
> cf.
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Sep/0283
>


B.2 datatypes

Peter agreed with my analysis that this is an error in S&AS, and that test
I5.8-016 should be correct.

Hence

PROPOSE that S&AS be modified to require a triple <xxx> rdf:type
rdfs:Datatype . in OWL DL for any user defined datatype <xxx>, even when
only used inside a typed literal.

This is no change to test and small change to S&AS.

B.1 uris as annotations

While we still disagree on this one, I am happy to concede - unless other WG
members feel strongly about this.
If someone else wishes to propose along the lines
[[
Modify resolution of Issue 5.26 OWL DL Syntax, so that uris used only as the
obejct of annotation properties do not need to have a type triple.
]]

This would be no change in S&AS and small change to test.

I would abstain.


Jeremy

Received on Tuesday, 28 October 2003 07:57:39 UTC