W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2007

Re: two test cases for OPTIONAL { { ... FILTER ( ... ) } } construct

From: Seaborne, Andy <andy.seaborne@hp.com>
Date: Wed, 03 Oct 2007 12:32:51 +0100
Message-ID: <47037DE3.9060805@hp.com>
To: Lee Feigenbaum <lee@thefigtrees.net>
CC: 'RDF Data Access Working Group' <public-rdf-dawg@w3.org>



Lee Feigenbaum wrote:
> As promised in the teleconference.
> 
> See the two tests
> 
> optional-filter/manifest#dawg-optional-filter-005-simplified
> 
> and
> 
> optional-filter/manifest#dawg-optional-filter-005-not-simplified
> 
> ...which differ only in the results.
> 
> I'm going to solicit implementation behavior and desire on this as 
> discussed in the teleconference.
> 
> FWIW, Glitter+Anzo passes the not-simplified test and fails the 
> simplified test -- I need to dig a bit but I think this is because I 
> attach the FILTERs to the OPTIONAL when I'm parsing, and even though I 
> do then simplify the group-containing-a-single-group construct, I don't 
> re-attach the FILTER at that point. I'd be content with either behavior, 
> especially if one is clearly preferred by implementations.
> 
> Lee
> 

ARQ passes #not-simplified (and fails #simplified) when no simplification is 
applied or if it is applied after the whole algebra expression has been 
created.  This is the current default behaviour.

It can also pass #not-simplified, and fail #simpified.  It is just a question 
of when the simplification is applied.

I prefer the separation of simplification from algebra creation (i.e 
simplification does not alter the outcome).  That is pass #not-simplifiedaand 
fail #simpified.

	Andy
Received on Wednesday, 3 October 2007 11:33:12 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:37 GMT