- From: Sandro Hawke <sandro@w3.org>
- Date: Mon, 01 Jun 2009 12:22:34 -0400
- To: "Seaborne, Andy" <andy.seaborne@hp.com>
- cc: "public-rdf-text@w3.org" <public-rdf-text@w3.org>
> > > http://lists.w3.org/Archives/Public/public-owl-comments/2009May/0009
> > >
> > > The issue about the results of FILTER functions, all algebra operators
> > > and how to pass constraints into a matching as some engines might (and
> > > do).
> > >
> > > Just saying "results" does not work. That only applies to what comes
> > > out in SPARQL results.
> > >
> > > We have three layers:
> > >
> > > 1 - Results formats (SPARQL XML Results or RDF graphs)
> > > 2 - Algebra and FILTER functions
> > > 3 - BGP matching.
> > >
> > > And also the query syntax (4).
> > >
> > > The text only covers (1) and (4). Change the matching and the correct
> > > behaviour at level 2 is undefined.
> >
> > It sounds like the solution is to extend the wording about syntaxes to
> > also cover interfaces between software sub-systems, and then perhaps
> > give one of these as an example. Would that do it?
>
> The reason this case is special is that it is an interaction with the
> SPARQL spec - it's not a private API matter.
>
> Therefore, I am suggesting explicit mention of extend BGP matching.
I've been staring at this, and I still don't get it. Can you put it in
programming terms for me? Maybe tell me what decisions you, as an
implementor of SPARQL system, expect to face because of this draft?
What is it that you think this draft is telling to do or not do to your
code, in order to remain conformant with all relevant W3C specs?
-- Sandro
Received on Monday, 1 June 2009 16:22:39 UTC