RE: draft response for LC comment 8 MS2

Hi!

Sorry for the late response. 

I think this draft makes a technically invalid suggestion, but this can
easily be corrected. See below!

>-----Original Message-----
>From: public-owl-wg-request@w3.org [mailto:public-owl-wg-request@w3.org]
>On Behalf Of Peter F. Patel-Schneider
>Sent: Wednesday, March 04, 2009 1:15 PM
>To: public-owl-wg@w3.org
>Subject: Re: draft response for LC comment 8 MS2
>
>[Draft Response for LC Comment 8:] MS2
>
>Dear Michael,
>
>Thank you for your message
>  <http://lists.w3.org/Archives/Public/public-owl-
>comments/2009Jan/0006.html>
>on the OWL 2 Web Ontology Language last call drafts.
>
>As far as the Direct Semantics of OWL 2 is concerned, a facet value
>could be any syntactic entity, 
              ^^^^^^^^^^^^^^^^

I think this is not valid. The elements < F v > of the 
facet space N_FS(DT) (for a given datatype DT) consist of 
an IRI F and a *(data) value* v. v it is not a syntactic 
entity. I will demonstrate this below.

>thus the wording "arbitrary object".  One
>could have a facet on integers called
>ex:greaterThanNumberOfOneBitsInUTF16Encoding, which takes arbitrary
>strings as values (e.g., the 12-character string "5"^^xsd:int) 

This is true for a datatype restriction expression in an ontology. 
However, the corresponding element of a facet space is a 
< facet , *value* > tuple of the form

  < ex:greaterThanNumberOfOneBitsInUTF16Encoding , 5 >

with the number 5 (a value), not with the literal "5"^^xsd:int 
(a syntactic entity).
 
>out those integers that were greater than the number of one bits in the
>UTF-16 encoding of the string.  One could also have a facet on decimal
>called ex:belongsToXSDDataType, which takes names of XSD data types
>derived from decimal as values (e.g., int) and picks out those values
>that belong to the value space of the XSD data type.
>
>One could even have a facet that takes arbitrary IRIs as values or even
>blank node identifiers.  However, a datatype map can only depend on the
>syntactic form of these facet values, not any of their properties in an
>ontology.

I'm not sure whether I understand this paragraph. But it might have to 
do with my original question. 

Imagine that datatype restrictions would 
allow arbitrary IRIs as the names for facet values. In OWL 2 DL, 
IRIs can only denote individuals in the object domain 
(same for bNodes). So this would mean that a tuple <F v> 
in a facet space can have object domain elements 
as their second component.

However, datatypes and datatype maps are defined independently 
from the interpretations which are "using" them. So nothing is 
known about the used object domain, when a datatype's facet space 
is defined. So there cannot really be datatypes that have facet spaces 
with individuals as the second component of a facet space entry,
since they would have to make assumptions about the content
of the object domain. 

So I think the second component of a facet space's element can only
be some data value, and nothing else.

This is where my original question came from. By "arbitrary object",
I wasn't sure whether it is meant that object domain elements are
allowed to be components of the elements of a facet space. 
I didn't believed that this makes any sense, hence my LC comment.

>The OWL 2 Functional Syntax restricts facet values to be literals, 
>and
>all the OWL 2 facets only depend on the data value that the literal
>corresponds to.  Therefore the working group has decided to modify the
>definition of datatype maps to require that facet values be literals, 
>as is already the case in the Functional Syntax. 

The Functional syntax can only refer to the value of a facet-value pair
by means of a literal. However, literals are part of a vocabulary. 
Datatypes and datatype maps do not have any notion of a literal.

Please compare the set V_FA in the definition of a "Vocabulary"
with the set N_FS in the definition of a datatype map (both
defined in Section 2.1 of the Direct Semantics). It is the set
V_FA(DT) which contains tuples of the form 

  < F , LITERAL >

where LITERAL is in fact a literal having the form LEXICALFORM^^DT_1 
(DT_1 does not need to equal DT !). 

The definition of V_FA(DT) states that the 
interpretation function LS (which is part of the datatype map!) 
has to map LEXICALFORM to a VALUE such that the tuple 
< F VALUE > exists in N_FS(DT). 

But according to the definition of LS, 
VALUE is a *data value*, being an element of (DT_1)^DT,
which is the value space of the data type DT_1 .

>The working group feels
>that this is an editorial change.

Yes, agreed. But the change should be

  "arbitrary object" --> "arbitrary (data) value".

>The diffs are:
>http://www.w3.org/2007/OWL/wiki/index.php?title=Syntax&diff=18711&oldid=
>18707
>The Direct Semantics document is not being changed, as permitting
>arbitrary objects there is not a problem.

The Direct Semantics really should be changed in the same way as the 
Structural Spec. The term "arbitrary object" should be replaced by 
"arbitrary (data) value". As argued above, nothing else than concrete 
data values can be the second component of a facet space's element.

>Please acknowledge receipt of this email to
><mailto:public-owl-comments@w3.org> (replying to this email should
>suffice). In your acknowledgment please let us know whether or not you
>are satisfied with the working group's response to your comment.
>
>Regards,
>Peter F. Patel-Schneider
>on behalf of the W3C OWL Working Group

Best,
Michael

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
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, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================