Re: comments on SPARQL QL, protocol, rf1, tests, requirements from outside the WG

On Apr 12, 2005, at 9:46 PM, Dan Connolly wrote:

> On Tue, 2005-04-12 at 21:03 -0400, Bijan Parsia wrote:
>> On Apr 12, 2005, at 4:50 PM, Dan Connolly wrote:
>>
>>>  In post-telcon discussion, I told Andy that I'd answered many of the
>>> comments, and I'd tell him which ones I have not answered. So here's 
>>> a
>>> list, in roughly newest-first order, back to about 23 March. I 
>>> presume
>>> the 31 March
>> [snip]
>>
>> I'm v. overtired, but I've seen no mention of the many issues I've
>> raised with regard to the formal bits, including the fact that I
>> believe, at the moment, that any bnodes in the query graph will make
>> the query fail.
>
> My message was about comments from outside the WG. You're
> inside the WG.

Oh yeah, I forget :)

>> Are these all considered editorial? If so, I'd like some mention of
>> that fact, and an indication that they've been dispatched to the
>> editors.
>
> Your comments are on the editors' todo list...
>
> "EP and AFS ack daveb's comments and bijan's"
>  -- http://www.w3.org/2005/04/12-dawg-irc
>
> We didn't discuss them in substance on the teleconference, but
> that doesn't mean I think they're strictly editorial. I'll keep
> an eye out for things that are substantive and conflict with earlier
> decisions, and schedule discussion of those.

Ok. If I have some notice, I can prioritize attending such a telecon to 
discuss them.

> For things that the WG hasn't decided, even if they're substantive,
> the editors are free to incorporate design changes.

Oh. Hmm. Interesting. Ok, that's different than other groups I've 
worked with. Cool.

>> I would also like to know the normative status of various bits of the
>> spec. Which trumps, definitions or main body text?
>
> The goal is that they're consistent. If they conflict, there's
> a bug. I'm disinclined to say, as a matter of policy, where
> the bug is.

Hmm. I prefer one specification, not two, if possible. I would prefer 
the formal stuff to trump (e.g., the grammar should trump examples or 
prose).

Cheers,
Bijan Parsia.

Received on Wednesday, 13 April 2005 01:52:33 UTC