Re: Importing SKOS model in ontology editors : problem with OWL full

Hi Sean,


I understand your point, and I was aware that there where some 
discussions. I'm personally very fond of SKOS, and I'm trying to show 
our partners it's benefits. Because of that I'm very disappointed with 
the results w.r.t. this OWL full issue.

We now have the situation that we have (a) a normal language 
recommendation and (b) an OWL1 full ontology, (c) intentionally no 
formal version which is tractable, e.g. OWL1 DLP , OWL2 EL, OWL2 RL or 
sth. similar.

Because an OWL1 full ontology is incomplete and not sound, from a semweb 
practicioner point of view this situation means that we simply have 
*nothing* at hand which can be used in professional semantic systems -- 
at least in systems which have to strive at soundness and completeness 
of their models.  SKOS 2008 doesn't fit practical needs; for the time 
being it's purely academic.


Alistair Miles wrote:
 > Just to let you know that last week the SKOS Reference Candidate
 > Recommendation was published, ...
 > The working group will shortly be sending out an official calling for
 > implementations, so watch this space.

With this OWL1 full issue it is *not possible* to provide a sound and
complete implementation. Are the editors aware of that?

It's also part of my job to write project proposals. Until now I have 
written "... We also strive at integrating SKOS terminologies into XYZ". 
If the SKOS editors insist not to define sth. like an OWL1 DL (or 
alternatively an OWL2) version, in the next proposal I'd have to write 
sth. like
     "Because inferencing with SKOS 2008 is not sound neither complete, 
we'll define a tractable subset of SKOS -- called 'skos lite' -- in 
order to allow for integrating terminologies similar to SKOS into XYZ."

Nobody want's to have such a situation.  What could be done?



yours,
Johannes




Sean Bechhofer wrote:
> 
> Bernard, Johannes
> 
> The OWL Full/DL issue was discussed in the WG (see, e.g. ISSUE-38 [1]). 
> During the LC period, the WG took a decision to change the typing of 
> labelling and documentation properties to be Annotation Properties (see 
> also discussion concerning resolution of ISSUE-157 [2]). Although the 
> presence of subproperty assertions between the labels and rdf:label and 
> and the documentation properties and skos:note violate the DL 
> constraints, it was our understanding that OWL 2 would move towards 
> supporting subproperty relationships between annotation properties. 
> Adopting this design will hopefully minimise potential changes in the 
> future in order to provide an OWL2-DL conformant vocabulary (personally, 
> I also believe that defining those properties as annotation properties 
> is also more appropriate).
> 
>> Here you point to such a prune candidate:
>> > http://www.w3.org/2006/07/SWD/SKOS/reference/20081001/skos-dl.rdf
> 
> I have regenerated the "DLised version" of the SKOS vocabulary [3]. The 
> version is produced by removing some of the axioms that violate the DL 
> constraints, in particular, those axioms relating to the subproperty 
> assertions concerning labelling and documentation properties. This is a 
> similar approach to that you discussed for the geonames. Note that this 
> does *not* form part of the SKOS recommendation, but may be of some use 
> in tools (it was originally done to support some tooling being developed 
> by Simon Jupp [4]).
> 
>> I have fed this to
>> - http://www.mygrid.org.uk/OWL/Validator
>> which gives back OWL DL plus some minor errors ("Possibly using wrong 
>> vocabulary (rdf:Property instead of owl:[Object|Data]Property)...")
> 
> These aren't errors, but are rather warnings. They are generated because 
> the SKOS vocabulary contains (redundant) assertions about properties. 
> For example, there are triples that state:
> 
> skos:inScheme rdf:type owl:ObjectProperty.
> skos:inScheme rdf:type rdf:Property.
> 
> These redundant triples were included to support non OWL aware 
> applications [5].
> 
>> Then we have a first question:
>> - Would the editors of SKOS agree that this (or a similar) OWL DL 
>> subset   reflects most of their *intended* semantic?
> 
> Speaking personally, yes (and I would hope so as I generated it :-). As 
> discussed above, it is missing the subproperty assertions (and other 
> statements concerned with metadata -- dc/dct vocabulary), but in so far 
> as it is possible to produce a SKOS OWL DL vocabulary, I believe it to 
> be a good fit.
> 
> Cheers,
> 
>     Sean
> 
> [1] http://www.w3.org/2006/07/SWD/track/issues/38
> [2] http://www.w3.org/2006/07/SWD/track/issues/157
> [3] http://www.w3.org/2006/07/SWD/SKOS/reference/20090315/skos-dl.rdf
> [4] http://code.google.com/p/skoseditor/
> [5] http://www.w3.org/2008/11/18-swd-minutes.html
> 
> -- 
> Sean Bechhofer
> School of Computer Science
> University of Manchester
> sean.bechhofer@manchester.ac.uk
> http://www.cs.manchester.ac.uk/people/bechhofer
> 
> 
> 
> 
> 


-- 
Dr. Johannes Busse, Senior Researcher
An der RaumFabrik 29, D-76227 Karlsruhe
Reg. Office: Karlsruhe, Amtsger. Mannheim, HRB 109540
Managing Directors:    Prof.Dr.J.Angele,  H.P.Schnurr
http://www.ontoprise.de   | phone x49(721) 509 809-62
mailto:busse@ontoprise.de | mobile x49(163) 509 80-62

Received on Wednesday, 25 March 2009 14:07:10 UTC