RE: ACTION-93 / ISSUE-63: Initiated work on OWL-1.1-Full semantics

>-----Original Message-----
>From: Peter F. Patel-Schneider [mailto:pfps@research.bell-labs.com] 
>Sent: Tuesday, March 25, 2008 8:00 PM
>To: Michael Schneider
>Cc: public-owl-wg@w3.org; jjc@hpl.hp.com
>Subject: Re: ACTION-93 / ISSUE-63: Initiated work on 
>OWL-1.1-Full semantics
>
>Looks good.

Many thanks, Peter! :)

>General points:
>
>1/ 
>
>IF AND ONLY IF
>	C
>THEN
>	D
>
>is not good grammar and is very hard to understand.  It should 
>either be 
>
>	C
>IF AND ONLY IF
>	D
>
>or
>
>IF 
>	C
>THEN
>	D
>
>depending on what is meant.

Alright! I have put TF-action 16 on myself to change this! :)

  <http://www.w3.org/2007/OWL/wiki/FullActions#Open_TF_Actions>

>A few specific points:
>
>1/ Axiomatic triple:
>
>	owl11:disjointObjectProperties rdf:type owl:ObjectProperty
>
>should be
>
>	owl11:disjointObjectProperties rdf:type rdf:Property
>
>Similarly for owl11:disjointDataProperties and owl:onProperty and
>rdfs:subPropertyOf and owl11:propertyChain and owl:unionOf, at 
>least.  The domain and
>range of rdfs:subPropertyOf are not owl:ObjectProperty.

You are right! From looking at sections 5.2 and 5.3 of the AS&S, only the
following seem to hold:
 
 CEXT_(S_I(owl:ObjectProperty)) = IOOP = P_I = CEXT_I(S_I(rdf:Property))

So owl:ObjectProperty and rdf:Property do not necessarily denote the same
property resource.

Hence, TF-action 17 on me to change this everywhere where needed!

Thanks!

>2/ I do not believe that a comprehension principle is needed for
>   owl11:disjointUnionOf (so long as there is a comprehension principle
>   for lists of descriptions) and the syntax requires that the
>   disjointUnion be a named class.

I don't understand your last remark about the requirement of a named class.

For your first point: AFAICT, this comprehension principle is redundant. If
I start from its premises: 

  (1a) l is a sequence of d_1, ..., d_n over IOC,
  (1b) CEXT_I(d_i) ^ CEXT_I(d_k) = {} for 1 <= i < k <= n

then I will get from (1a) plus the comprehension principle for owl:unionOf
(1.0-Full semantics):

  (1c) EXISTS c in IOC: <c,l> in EXT_I(S_I(owl:unionOf))

So I receive from the semantic condition of owl:unionOf (1.0-Full semantics)
that for this class c holds:

  (1d) CEXT_I(c) = CEXT_I(d_1) u … u CEXT_I(d_n)

(1d) + (1b) plus the main semantic condition for owl11:disjointUnionOf now
leads to

  (1e) <c,l> in EXT_I(S_I(owl11:disjointUnionOf))

So we could drop the comprehension principle for disjointUnionOf for
technical reasons. However, having it doesn't seem to hurt (as long as there
is also one for owl:unionOf). And I prefer having a common style of
semantics for analog axioms. And I don't like to answer questions of the
kind: "Hey, did you forget to define the comprehension principle?". :) So I
would like to keep it in for "symmetry reasons". What do you think?

>3/ The main semantic conditions for negative property 
>assertions have an
>   unbound variable.   This is probably a bad thing in an IFF 
>condition.
>   I suggest instead using an IF condition and a comprehension 
>principle
>   (like for the n-ary constructs).

Argh, you are right! TF-Action 18 on fixing this. Thanks! 

(BTW, I am not the first one who made this error... leading to fatal
consequenses! I call this the "prenexity trap". :))

>4/ I believe that owl:subPropertyOf should be rdfs:subPropertyOf

Of course, thanks! Ok, action 19, because I have to look through all the
pages, and I am tired now.

>5/ The RDFS-compatible semantics for OWL 1.0 would be incorrect if
>	owl:FunctionalProperty rdfs:subClassOf owl:ObjectProperty 
>
>   was an axiomatic triple.  Similarly for the domain and range of
>   owl:sameAs and owl:differentFrom.  Similarly for the domain 
>and range
>   of owl:equivalentProperty and the range of owl:onProperty and
>   owl:hasValue.

This seems closely related to the point above, so covered by action 17.
Thanks, anyway.

>peter
>
>PS: Good catch on owl:AllDifferent

Thank you! But, I'm afraid, this was only the easily visible tip of the
iceberg... :-]

Here are a few question, which I are very important for me to know.

(A) How do you think about my approach to introduce the "missing" axiomatic
triples? Was is intended to not include them, and if so, for what reason? Or
has it just been forgotton?

(B) Why are there only "ONLY-IF" semantic conditions for restrictions? The
introduction of sec. 5.2 in the AS&S says:

  "The only-if semantic conditions for the [restrictions] are needed 
  to prevent semantic paradoxes and other problems with the semantics."

Can you please elaborate on this?

Many thanks again for your (very fast, wow!) help!

Cheers,
Michael

--
Dipl.-Inform. Michael Schneider
FZI Forschungszentrum Informatik Karlsruhe
Abtl. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: Michael.Schneider@fzi.de
Web  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555

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