RE: Suggestion to change RDF encoding of HasKey axioms [RE: I've implemented the change to the syntax of HasKey]

>-----Original Message-----
>From: Ivan Herman [mailto:ivan@w3.org]

>So, just as a thought experiment: why not say that, by default,
>properties listed in a hasKey are object properties (ie, I can put there
>a list of properties, simply), but I can have a separate list which
>explicitly lists datatype properties? Ie,
>
>- if I want to have FS-compatible encoding, I would separate these two
>and make the encoding proper in FS
>- if I do not really care about FS, I would probably disregard the
>datatype property part of keys and just use the list of properties.
>
>Ivan

This would not properly work. 

Say the combination of knowing some person and the date since knowing him would make a key. You might encode this as

  (1) _:x rdf:type owl:Key .
  (2) _:x owl:keyClass foaf:Person .
  (3) _:x owl:keyObjectProperties ( foaf:knows ) .
  (4) _:x owl:keyDataProperties ( ex:since ) .

Now, let's say you have the following data:

  ex:alice rdf:type foaf:Person ;
           foaf:knows ex:charly ;
           ex:since "2003-08-17"^^xsd:date .

  ex:bob rdf:type foaf:Person ;
         foaf:knows ex:charly ;
         ex:since "1994-04-12"^^xsd:date .

With your approach of an *optional* data property list, you'll be surprised to learn that

  ex:alice owl:sameAs ex:bob .

Why? Because you would have actually created *two* key axioms. The triples

  (1), (2), (3)

would also make a valid encoding for a key axiom, stating that two foaf:PersonS are the same, if they foaf:knows the same person.

We had a similar problem in the past with the RDF encoding for QCRs, when they were encoded just as cardinality restrictions with an additional owl:onClass property. Doing such "monotonic extension" of existing RDF encodings will generally call for trouble.

Cheers,
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
=======================================================================