Re: significant problem for moving OWL to Last Call

From: <herman.ter.horst@philips.com>
Date: Thu, 27 Mar 2003 17:12:39 +0100
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>

Message-ID: <OFC873D1CD.AEA908BC-ONC1256CF6.0052B052-C1256CF6.005934BF@diamond.philips.com>
```
last call version of the RDF Semantics document, I know that
Pat sees IC and ICEXT as add-ons, just as Peter wants them to see
(and as I also want them to see).

Note that in the text Peter cites,

>       An rdfs-interpretation of V is an rdf-interpretation
>       I of (V union rdfV union rdfsV) *with a distinguished subset IC
>       of the universe and a mapping ICEXT from IC to the set of
>       subsets of IR*, which ...

IC and ICEXT are not put *into* the tuple that forms I.
I believe they are listed here as add-ons.

But the possibility of such confusion should be excluded.

The following modification of the text that Peter proposed below
does not only exclude this possibility of confusion, but also
seems to keep the intent expressed in the current editor's version
of the RDF Semantics document.

An rdfs-interpretation of V is an rdf-interpretation I of (V union
rdfV union rdfsV) which satisfies the following semantic
conditions
and all the triples in the subsequent table, called the RDFS
axiomatic triples.  For convenience, and to make the semantic
conditions easier to understand,
the set of classes IC is defined as
IC = { y | <y,I(rdfs:Class)> is in IEXT(I(rdf:type)) }
and the function ICEXT from IC into the powerset of IR is
defined, for each x in IC, as
ICEXT(x) = { y | <y,x> is in IEXT(I(rdf:type)) }.
This implies that
IC = ICEXT(I(rdfs:Class)).

In agreement with Peter, I would then also remove the other
discussion of IC and ICEXT before the definition of an
rdfs-interpretation.
And I would remove not only the second but also the first condition
from the table (which deals with ICEXT and becomes clearly also
superfluous).

Moreover, I believe that in [1] and [2] I listed change suggestions
that make the S&AS document completely consistent with this
version of the RDFS semantics .
([2] is later in the same thread, and contains a copy of the
essential part of [1].)

Herman

[1] http://lists.w3.org/Archives/Public/www-webont-wg/2003Mar/0209.html
[2] http://lists.w3.org/Archives/Public/www-webont-wg/2003Mar/0232.html

>Herman ter Horst just brought to my attention a significant change to the
>RDFS semantics in the editor's version of the RDF semantics document.
>
>This change involves adding new constructs to the definition of RDFS
>interpretations as follows:
>
>       An rdfs-interpretation of V is an rdf-interpretation
>       I of (V union rdfV union rdfsV) *with a distinguished subset IC
>       of the universe and a mapping ICEXT from IC to the set of
>       subsets of IR*, which satisfies the following semantic
>       conditions and all the triples in the subsequent table,
>       called the RDFS axiomatic triples.  [Emphasis added]
>
>Previously RDFS interpretations used the same structure as RDF
>interpretations, and IC and ICEXT were conveniences only.
>
>To track this change will require significant changes to S&AS.  I do not
>feel that OWL can go to last call without some resolution of this new
>issue.
>
>Peter F. Patel-Schneider
>Bell Labs Research
>Lucent Technologies
>
>PS:  If I had my druthers, I would change RDFS interpretations back to
>being just like RDF interpretations as they were in the Last Call version
>of the RDF semantics.  If a clarification is needed I would proceed
>somewhat along the lines of defining rdfs-interpretations as
>
>       An rdfs-interpretation of V is an rdf-interpretation I of (V union
>       rdfV union rdfsV) which satisfies the following semantic
conditions
>       and all the triples in the subsequent table, called the RDFS
>       axiomatic triples.  For convenience, and to make the semantic
>       conditions easier to understand, ICEXT is defined as
>               ICEXT(x) = { y | <y,x> is in IEXT(I(rdf:type)) }
>       and IC is defined as
>               IC = { y | <y,I(rdfs:Class)> is in IEXT(I(rdf:type)) }
>       which is the same as saying
>               IC = ICEXT(I(rdfs:Class))
>
>and then removing the second semantic condition.  I would also remove the
>discussion of IC and ICEXT before the definition of an
rdfs-interpretation,
>but not the discussion of a class, although minor changes would need to
be