LIMIT clause in right place?

In:
   http://www.w3.org/TR/2005/WD-rdf-sparql-query-20050217/

I'm unconvinced that it is appropriate for a LIMIT clause to be part of the 
query language.  Among other things, it apparently has the effect of making 
a query indeterminate against some given source data, which doesn't seem to 
be a desirable property.  Also, IIRC, there is a functional overlap here 
with the query protocol, which also specifies a result limit.

Limiting the number of results would appear to be a purely operational 
matter of optimizing resource usage, and as such it seems to me that 
limiting the result set is more appropriately dealt with as part of the 
protol.  Also, when limiting the number of results for a request, I see no 
way to ask for the "next n", which I think is a commom requirement, and one 
that I think is much more easily accommodated in the query protocol than in 
the query language.

#g


------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact

Received on Friday, 8 April 2005 11:49:02 UTC