Re: Comment: SPARQL Query conformance section

On Mon, 2005-10-24 at 12:28 -0400, Eric Prud'hommeaux wrote:
> On Mon, Oct 24, 2005 at 09:31:40AM -0500, Dan Connolly wrote:
> > On Mon, 2005-10-24 at 09:38 -0400, Eric Prud'hommeaux wrote:
> > > We have a coment asking for a conformancy section in SPARQL Query:
> > > http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Jul/0027
> > > 
> > > DanC, you hvae argued that we should leave conformance to the protocol
> > > spec (or an argument near that).
> > 
> > I've argued that the QL spec should specify conformance of documents
> > and answers to the QL syntax and semantics,
> > and that the protocol spec should specify conformance to the protocol.
> 
> Roger that. I'm trying to sort out our task here. You specifically
> requested that we not include RFC2119 (MAY MUST MIGHT) or include the
> word "implementation".

Right; a language can and should be specified without reference
to implementations.

And I think it's silly to say that "the word SELECT MUST be
followed by a space"; it's like saying "2 + 2 MUST be 4".
Better to just say "2 + 2 is 4", or in our case, as
you point out below, "A SPARQL query string is a sequence
of characters in the language defined by the following grammar. ...".


>  I was persuaded by this that we could define a
> language and that conformance could come from the protocol document.

No, conformance of strings to the language definition is in scope,
as is conformance of answers to queries. Our last call decision
was a decision that we had specified those to our satisfaction.

The commentor seems to prefer that we organize that material
differently.

> If the plan is different, please advise. If we are expected to provide
> a conformance section without MAY and MUST or talking about
> implementations, I would like some help. I thought that this was the
> conformance plan
>   <http://www.w3.org/TR/rdf-sparql-query/#grammar>:
> [[
> A SPARQL query string is a sequence of characters in the language
> defined by the following grammar, starting with the Query production.
> ]]
> Maybe s/query string/query/ and add?:
> [[
> The semantics of every SPARQL query are defined in this document. 
> ]]

Yes, that's the relevant current text.

The commentor is asking for a different organization. That's
an editorial matter. You will please make an editorial judgement
and either
  (a) try to satisfy the commentor with the current text, or
  (b) propose different text in mail to the WG;
   (1) if you think the different text doesn't involve any design
   changes, you can ask that I confirm that, so that you can
   respond to the commentor
   (2) if you come up with design changes while you're at it,
   we'll need to make a WG decision.

The commentor might be satisfied by something like changing the
Grammar section heading to "Grammar and Query String Conformance".

> > >  I think the WG has tacitly or
> > > activley agreed. Could you send mail to the commentor arguing your
> > > position?
> > 
> > No; I think the comment is editorial; it's asking for a different
> > form of emphasis on the conformance stuff. I leave that to
> > the editors.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Monday, 24 October 2005 16:51:55 UTC