W3C home > Mailing lists > Public > semantic-web@w3.org > June 2010

RE: An idea I need help with, or told to stop wasting time on!

From: Michael Schneider <schneid@fzi.de>
Date: Mon, 7 Jun 2010 09:39:24 +0200
Message-ID: <0EF30CAA69519C4CB91D01481AEA06A001E99786@judith.fzi.de>
To: <nathan@webr3.org>
Cc: "Linked Data community" <public-lod@w3.org>, <semantic-web@w3.org>
>-----Original Message-----
>From: semantic-web-request@w3.org [mailto:semantic-web-request@w3.org]
>On Behalf Of Nathan
>Sent: Sunday, June 06, 2010 11:51 PM
>To: Michael Schneider
>Cc: Linked Data community; semantic-web@w3.org
>Subject: Re: An idea I need help with, or told to stop wasting time on!
>
>Michael Schneider wrote:
>> Hi!
>>
>> Just a few notes concerning your ideas and OWL DL (I don't know
>whether this
>> is important for you or not, but some people might find it relevant):
>
>Thanks Michael,
>
>Very useful and indeed relevant (thanks!).
>
>To summarise, everything mentioned so far by me is fine in OWL Full and
>RDF(s), but not in OWL DL.

I should have mentioned that you could make ex:value an
owl:AnnotationProperty, which would allow you to have all of URIs, literals
and bnodes in object position. But this, of course, has other drawbacks in
OWL DL, apart from not looking very justified conceptually (ex:value is
probably not meant as a means to add comments to a resource?). If you make
it an annotation property, most OWL constructs cannot be used with the
property anymore. For example, it may make sense to state that ex:value is a
functional property, or to put a has-value restriction on it in some
scenarios. That's all not possible then anymore. In OWL 2 DL, you could at
least put a range axiom on it, but it would not have any semantic
consequences, i.e. an OWL DL reasoner would completely ignore both the
property and the axiom on it. This may lead to surprises.

So, my general view is that making a property, which is not naturally sort
of a commenting property (such as rdfs:comment), an annotation property is
only acceptable, if you exactly know what you are doing and if you have full
control over the property's use. If you expect to publish the property to be
used by others, and if there are possible scenarios where one might like to
use the property in an OWL construct (e.g. an axiom) or even do reasoning
with it, then don't make it an annotation property.

>Thus is it safe to say that this would be a problem in OWL DL as well?:
>
>   :x owl:sameAs 'a literal'^^xsd:string .

No, owl:sameAs cannot be used with literals in OWL DL. It can only be used
with URIs (named individuals). What you are doing here is, again, "genuine"
OWL Full, because OWL Full treats data values as individuals.

>And I guess the take-away is, that if one was to go for something as
>described in the original post, it would not be OWL DL compliant.

Consider SKOS-XL [1]:

  ex:foo skosxl:prefLabel [
      rdf:type skosxl:Label ;
      skosxl:literalForm "foo" ]

Here, skosxl:prefLabel is specified as an object property [sic!] and
skosxl:literalForm is a data property (while the better known skos:label
property is an annotation property). This works in DL, but only if you use
those properties as a "team". In your original example, you have used
foaf:name, which was a data property, and this does not work. Also, you
cannot use skosxl:literalForm with a URI as an object, what you did with
ex:value in your earlier post. So, you can do it in DL, but you don't have
very much usage freedom. Thus, check your use cases! 

>ps: If I get to the stage of trying to express any of this in an OWL
>ontology (FULL I guess!), would it be okay to send through to cast your
>eye over.

Feel free. For OWL Full, actually, it's as simple as this: syntactically, if
it is in RDF (and it always is), then you are in OWL Full (a no-brainer),
simply since the syntax of OWL Full is defined to be (unrestricted) RDF. And
if you want to do OWL Full-style reasoning, then the OWL 2 RL/RDF Rules
language [2] and corresponding reasoners (e.g. [3]) are often sufficient
(though sometimes not, depends on your usecases).

>Many Regards,
>
>Nathan

Best,
Michael

[1] http://www.w3.org/TR/skos-reference/#xl
[2]
http://www.w3.org/TR/2009/REC-owl2-profiles-20091027/#Reasoning_in_OWL_2_RL_
and_RDF_Graphs_using_Rules
[3] http://www.ivan-herman.net/Misc/2008/owlrl/

--
Dipl.-Inform. Michael Schneider
Research Scientist, 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
=======================================================================
Received on Monday, 7 June 2010 07:40:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:36 GMT