Re: [WN, SPARQL, ALL] Describe

McBride, Brian wrote:

>Reading the message I just sent (after our Christmas lunch, hic :) I'm
>reminded of my guidance to self, that we should couch our answers in
>terms of specific requirements of the work of SWBPD, i.e. are there
>specific guarantees we would like to see for DESCRIBE or Wordnet or SKOS
>resources.
>
>  
>
I, and I assume most others in this WG, have only practical experience 
with RDF query languages of which
a) the behavior is independent of the  vocabulary being queried
b) the nature of the query result is determinstic and specified in the 
documentation

IF
 there are currently SPARQL or other implementations that implement 
DESCRIBE, AND
 these implementations return something reasonable in the general case, AND
 these implementations are easy to configure to return something useful 
in the specific (e.g. WordNet) case
THEN
this WG might try to develop some best practices
ELSE
it will be hard to say much in favor of  DESCRIBE strictly in terms of 
best practices.

I have no major problems with DESCRIBE, I have problems with it as an 
integral part of the core of the first version of the first common 
semantic web query language ever.  If it is in a separate language or in 
the SPARQL extension profile or in SPARQL Level 3 or something the story 
would be different.

Jacco

Received on Thursday, 15 December 2005 15:42:20 UTC