Re: SKOS: COMMENT: Re: [Mulgara-general] Rules and SKOS

Hi David,

Thanks for forwarding this, this is great. This is also very timely,
as we're within the last call comment period, so now is the time to
speak if this work raises any issues.

There is one issue which I would particularly appreciate feedback
on. Looking at http://mulgara.org/trac/wiki/SKOS I see the following:

""" 
There is also a workaround in this definition file to deal with a
peculiarity in SKOS.

SKOS declares that skos:broader is a sub-property of
skos:broaderTransitive. Similarly, skos:narrower is a sub-property of
skos:narrowerTransitive. However, the documentation also declares that
"sub-property" does not inherit transitivity. This means that
skos:broader and skos:narrower do not have a rdfs:subPropertyOf
relationship to skos:broaderTransitive and skos:narrowerTransitive.
"""

Do you believe that the SKOS Reference misuses or misunderstands the
notion of "sub-property" as defined in RDFS and OWL? If so, then that
is a serious issue for us.

The SKOS Reference defines the SKOS data model as an OWL Full
ontology, with the standard OWL interpretations of sub-property and
transitivity. As I understand RDFS and OWL, transitivity is not
inherited by sub-properties of a transitive property. For example, the
ancestor relationship is transitive, but sub-properties of ancestor
such as father or grandmother are not transitive. Am I missing
something?

"""
To get around this, we've directly sub-classed
skos:broader/skos:narrower off skos:semanticRelation, and removed
their relationship to
skos:broaderTransitive/skos:narrowerTransitive. We have also added in
the explicit relationship:

skos:broaderTransitive(A,B) :- skos:broader(A,B).
"""

Again, based on my understanding of sub-properties and transitivity in
OWL, I do not see why this workaround is necessary. If I've
misunderstood, please do explain ASAP. If SKOS requires such a
workaround then we should change it.

Thanks again, and please do respond soon.

Alistair.


On Tue, Sep 30, 2008 at 07:53:52PM +0100, Simon Spero wrote:
>
> This is very interesting.
>
> This is also a very good example of just how the introduction of  
> broaderTransitive and narrowerTransitive   may, on further reflection,  
> turn out to be, upon further review, differently  weighted in the  
> balancing of factors than, on ones initial impression, may have seemed, 
> or rather, the converse.
>
> Or, to quote Professor Jane Greenberg, SKOS is neither Simple, or a  
> K.O.S.
>
> Simon
>
>
> On Sep 29, 2008, at 1:19 PM, David Wood wrote:
>
>>
>> This is awesome, Paul!  Thanks!
>>
>> I'm copying the SKOS list so they know about it.
>>
>> Regards,
>> Dave
>>
>> On Sep 28, 2008, at 7:26 PM, Paul Gearon wrote:
>>
>>> I've recently had to do a little bit of work with SKOS and Rules, and
>>> I thought it might be worth pointing out to people.
>>>
>>> You'll find a description on how to use rules in the Wiki at:
>>> http://mulgara.org/trac/wiki/Rules
>>>
>>> I've also included a file for partial SKOS entailments (partial,
>>> because I don't have collection support yet). This is attached to and
>>> described in:
>>> http://mulgara.org/trac/wiki/SKOS
>>>
>>> The SKOS rules may be informative for anyone wanting to use SKOS, or
>>> rules in general. Note that almost all the work is done in the 6  
>>> rules
>>> at the bottom of the file. Everything else is done with OWL
>>> vocabulary.  (There are 4 extra rules, some of which may be  
>>> redundant,
>>> plus a 5th that is necessary due to a strange requirement in SKOS,  
>>> but
>>> each of these are exceptional, rather than the norm).
>>>
>>> Regards,
>>> Paul
>>> _______________________________________________
>>> Mulgara-general mailing list
>>> Mulgara-general@mulgara.org
>>> http://mulgara.org/mailman/listinfo/mulgara-general
>>
>>
>
>

-- 
Alistair Miles
Senior Computing Officer
Image Bioinformatics Research Group
Department of Zoology
The Tinbergen Building
University of Oxford
South Parks Road
Oxford
OX1 3PS
United Kingdom
Web: http://purl.org/net/aliman
Email: alistair.miles@zoo.ox.ac.uk
Tel: +44 (0)1865 281993

Received on Wednesday, 1 October 2008 11:00:48 UTC