RE: Question re: HasKey entailments

>From: Jim Hendler
>Sent: Tuesday, December 30, 2008 10:44 PM

>Again, to be clear, I understand he design decision,  and I can see
>where this could be a useful property to have, but the traditional use
>of keys in databases (or inverseFunctional datatype properties)
>doesn't assume this sort of localness, 

Hm, of course, there are not really classes in RDBs, it's a different kind
of model. But keys normally belong to tables, aka collections of datasets.
Can't we say that the tables correspond to classes, and regard their
datasets as their (concrete) instances? Under this perspective, there can be
keys with the same name, e.g. "id", for two tables (classes) T1 and T2. But
if we have the data value "42" both in columns T1.id and T2.id, then the two
datasets (instances) referenced by these two keys will generally not be the
same. So the keys are really local to the respective "class" for which they
have been defined. They are at least not global for the whole database,
which would roughly correspond to the whole ontology in OWL.

Apart from this somewhat fancy consideration, there is a different aspect of
Keys in OWL 2: They also provide for compound keys, and this is really an
important RDB feature, and isn't available by inverse functional properties.

 
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