RE: SKOS Core second review

At 15:15 3/08/2005, Houghton,Andrew wrote:
>A thesaurus management system could be built using SKOS as the underlying 
>model,
>so I don't necessarily agree that it wouldn't be good practice to 
>produce/maintain
>a thesaurus in SKOS.

Yes, you could build a thesaurus management system (TMS) using SKOS as the 
underlying data model. However-- and I don't know why it always has to be 
me that points this out ;-( -- but a TMS built on SKOS would not comply 
with BS8723, since there is no way to associate data elements such as 
history note, source, definition, editorial note, etc. with a 
non-descriptor/non-preferred term. While the last draft of the revised ANSI 
standard is largely silent on this point, Figure 18 clearly shows source 
information being recorded about synonyms for a candidate term in AAT, so 
this is not just a British invention ;-).

As Mark points out, this doesn't mean that SKOS isn't very good for other 
things like interoperability. However I can say that there are simple, 
practical and (in some environments at least) compelling reasons why you 
would want to record and manage this kind of information in a TMS.

Ron

Ron Davies
Information and documentation systems consultant
Av. Baden-Powell 1  Bte 2, 1200 Brussels, Belgium
Email:  ron(at)rondavies.be
Tel:    +32 (0)2 770 33 51
GSM:    +32 (0)484 502 393 

Received on Wednesday, 3 August 2005 18:42:21 UTC