W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > April to June 2005

sort issue ready to close, right? [was: more sorting test cases]

From: Dan Connolly <connolly@w3.org>
Date: Mon, 09 May 2005 16:56:45 -0500
To: andy.seaborne@hp.com
Cc: public-rdf-dawg@w3.org
Message-Id: <1115675805.13402.56.camel@localhost>

In case anybody missed it under that subject line...

On Tue, 2005-05-03 at 17:34 +0100, Seaborne, Andy wrote:
> I propose the tests in
http://www.w3.org/2001/sw/DataAccess/tests/data/sort/ 
> address the issue and should be adopted.

I think we're ready to close this issue. If WG members think
otherwise, let them speak before 16 May 14:30Z (in
time for the 17 May agenda) or forever hold their peace.

The specification in

  http://www.w3.org/2001/sw/DataAccess/rq23/#solutionsResults
  $Revision: 1.328 $ of $Date: 2005/05/03 12:34:36 $

was good enough for Jeen to write these tests, and Andy
hasn't found any reason to reconsider after implementing
it. Andy could be reading his own mind, rather than the text,
of course.

I see some ToDos added in
1.327        (aseaborn 29-Apr-05)

<p class="todo">@@TODO@@Link to solution sequence modifiers</p>
<p class="todo">@@TODO@@Abstract model: pattern match, sequnce of
result form.</p>

and ideally I might like those ToDos to-done before we close
the issue... I'd like to study some formal definitions. Like
everybody else in the WG, if I thought they were sufficiently
important to have, I could have proposed some text myself.
I think there's been sufficient discussion of this design that
the risk is acceptable. i.e. the risk that filling in these
details will cause us to reconsider the design.

There seems to be some sentiment for postponing this issue.
I'm quite sympathetic to that, but we seem to have done much
of the work that postponing would have saved us... testing
and implementation, anyway. And postponing could have cost
us a lot of time in explaining why we don't have ORDER BY.
I'm still open to the proposal that sorting should be postponed,
but I'll probably defer to EricP on this one when it comes
time to decide. Last time I checked, there were likely
objections to postponing.

Is anybody likely to object to a proposal that the 1.328
spec and the tests/data/sort/ tests address the sort issue?
(with an action on somebody... probably Andy... to
finish the formal definitions)

If your position is that you haven't studied the design or
you don't think you have time to implement it in the next
few months, I think that's fine grounds to abstain, but
not very good grounds to object. Note that objections
only need to be reported to The Director if they include
"technical arguments and propose changes that would remove the
objection".
http://www.w3.org/2004/02/Process-20040205/policies.html#WGArchiveMinorityViews


If your position is that SPARQL should not have sorting
at all or should not have had it in this version, you
should have objected to our decision to add a sort objective.
http://www.w3.org/TR/rdf-dawg-uc/#d4.11


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E
Received on Monday, 9 May 2005 21:56:49 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:23 GMT