Re: thesaurusRepresentation-11

Hi all,

I had a chat with Dan at ESWC'06 and he suggested a solution: have a 
"fat" SKOS version that has the Term class, and an automated procedure 
that produces a "lean" version.

The fat version would have a class Term with a label property attached 
to it (and other stuff attached to it). The label property of the terms 
in a specific thesaurus can be automatically converted to skos:prefLabel 
and skos:altLabel properties on skos:Concepts to produce a lean version 
(i.e.  what the current SKOS produces).

Come to think of it, this is very similar to the approach we took with 
WordNet [1]. The Full version has Synsets, WordSenses and Words, where 
the final literals are in a wn:lexicalLabel attached to the Words. The 
Basic version consists of the Synsets plus additional file with 
wn:senseLabels. These are attached to Synsets, and their contents is 
from the lexicalLabels of the Words that are contained in the Synset.

Cheers,
Mark.

[1]http://www.w3.org/2001/sw/BestPractices/WNET/wn-conversion

Alistair Miles wrote:
> 
> Hi all,
> 
> I've added this item to the proposals and issues list, see:
> 
> http://www.w3.org/2004/02/skos/core/proposals#thesaurusRepresentation-11
> 
> I've included a link to this thread, so that Stella's comments are 
> noted. I included the point about "USE x OR y"/"UFO" because it will 
> probably interact with SKOS representation of "USE x + y"/"UF+", i.e. 
> the representation will need to clearly distinguish between the two, 
> even if the former is not recommended.
> 
> Cheers,
> 
> Al.
> 
> Stella Dextre Clarke wrote:
>> Mark,
>> Re point 1, yes, I dare say there may be other uses. Futile to try and
>> enumerate all of them! And I agree it is an advantage if SKOS is able to
>> accommodate all of the information carried in a thesaurus. That said, it
>> is possible to comply with ISO 2788 or BS 8723 and still have some extra
>> features that are not used anywhere else. To be able to accommodate such
>> extra features, you would need to have a capability for defining custom
>> relationships and data types. I vote for coping with what is described
>> in the standards, and what is commonly encountered, before tackling the
>> obscure  curiosities that may still be emerging.
>>
>> RE point 3, USE...OR/UFO  is not very common in my experience. There
>> tend to be few instances, even in those thesauri that do use it. But
>> exceptions are sure to occur somewhere. WE just don't recommend it, in
>> the standards.
>> Stella
>>
>> *****************************************************
>> Stella Dextre Clarke
>> Information Consultant
>> Luke House, West Hendred, Wantage, Oxon, OX12 8RR, UK
>> Tel: 01235-833-298
>> Fax: 01235-863-298
>> SDClarke@LukeHouse.demon.co.uk
>> *****************************************************
>>
>>
>>
>> -----Original Message-----
>> From: Mark van Assem [mailto:mark@few.vu.nl] Sent: 09 June 2006 08:16
>> To: Stella Dextre Clarke
>> Cc: 'Alistair Miles'; public-esw-thes@w3.org
>> Subject: Re: thesaurusRepresentation-11
>>
>>
>> Hi Stella,
>>
>> Some quick comments on your comments ;-)
>>
>>> 1. "Any type of annotation associated with a non-descriptor" Yes, I 
>>> believe this is useful and needed by a lot of vocabulary managers. I 
>>> don't see it as a requirement while using a thesaurus in connection 
>>> with searching a database (or other information resource), but it 
>>> certainly has uses while managing and maintaining the vocabulary.
>>
>> There may be other uses like for (semi-)automated mapping and NLP. More
>> general: It would be a pity if some thesauri get two RDF versions, one
>> in SKOS and one in another schema for the stuff that did not fit into
>> SKOS. E.g. how to deal with sources that are annotated with the other
>> version but should be used together?
>>
>>> 3. "1:n relationships between non-descriptors and descriptors ("USE x
>>> OR y"/"USO")"
>>
>> Would you have an idea on how often this relationship is used?
>>
>> Cheers,
>> Mark.
>>
>>
> 

-- 
  Mark F.J. van Assem - Vrije Universiteit Amsterdam
        markREMOVE@cs.vu.nl - http://www.cs.vu.nl/~mark

Received on Friday, 16 June 2006 10:41:46 UTC