Re: [SKOS] The return of ISSUE-44 (was Re: TR : SKOS Reference Editor's Draft 23 December 2007)

Hi Simon,
>
> On Jan 10, 2008, at 9:46 AM, Antoine Isaac wrote:
>> I'm sorry but I don't have it...
>> I don't have ISO available right now, but I will check it.
>> As far as Z39.19 is concerned I cannot see a reference to something 
>> like transitivity (but I might have overlooked, I just read quickly 
>> the section on se;antic relations)
>> How about the following example:
>> mountains regions BTI Himalaya
>> Himalaya BTP Everest
>>
>> Can we naturaly have Everest as a narrower term of montains regions? 
>> Does Z39.19 explicitly forbid that?
>
> I can't remember if Z39.19 explicitly prohibits this, but my personal 
> belief is that it BTI and BTP are distinct relationships.
>
> The logical test for   BT = BTG + BTP could be described as "is or is 
> part of a ";  Everest "is or is part of a " Mountain regions"  
> undoubtedly holds

Yes, but a true transitive BT that gathers all your possible 
interpretations (BTG.BTP is only one combination) should then lead to a 
logical test that tests the whether the link matches one among all the 
possible combinations of "is part of" and "is a" ("A is a part of B 
which is a C which is a part of a part of a D" and so on).
I don't think you would thus obtain something that your standards still 
recommend. I guess that actually your "is or is part of a " already 
falls out of what is considered good practice for thesaurus engineering. 
Andof course departs from the "precise meaning" that you advocate for 
broader.

Antoine

Received on Thursday, 10 January 2008 13:20:33 UTC