W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2006

Re: bnodeification (was: Re: SPARQL semantics: open issues for basic query patterns)

From: Pat Hayes <phayes@ihmc.us>
Date: Tue, 10 Jan 2006 10:15:07 -0600
Message-Id: <p06230900bfe987e509b7@[]>
To: Souripriya Das <souripriya.das@oracle.com>
Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>, bparsia@isr.umd.edu, Enrico Franconi <franconi@inf.unibz.it>, Sergio Tessaris <tessaris@inf.unibz.it>

>Quoting from Pat Hayes's e-mail 
>"Another way to define S(Q) is to imagine that every bnode in Q is 
>replaced by a distinct query variable (which cannot be SELECTed, 
>however) and then use this definition. There are cases in which this 
>definition is slightly tighter than the previous one, eliminating 
>some redundancy. Also this doesn't need the ordered-merge idea."
>This seems to be a simpler way. Has this been discussed already?

As I (dimly) recall, we did at one point consider not allowing bnodes 
in queries but this idea was rejected as too restrictive for the 
query writers (imagine taking some RDF and plopping variables where 
you want answers, but leaving the other bnodes alone, why not?) 
However, I think if we had realized the tar-pit we were getting into 
by allowing bnodes scoped to queries, we might have considered it 
more actively. It sure does make all the definitions easier (and I 
think conceptually clearer) if we assume that bnodes in queries are 
treated as simply a species of variable, as this neatly solves all 
the awkward 'scoping' problems that arise.


>- Souri.
>Attachment converted: betelguese2:souripriya.das.vcf (TEXT/Hdra) (00226A23)

IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Tuesday, 10 January 2006 16:15:35 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:00:50 UTC