Re: Querying multipl sources objective

Jim,

I would prefer to drop both 4.5 and 4.5.1 and concentrate point-to-point 
query (that allows for named, predefined intergrated graphs).

The UCs of asking Aggregate Query could be done by the client library (not 
perfected optimal but fairly close) and so we could be said to achieve it 
without anything in the recommendation.  I do want a recommendation where 
a server does not have to do anything to deal with multiple targets [*] 
and still be compliant with the full recommednation.  I am not in favour 
of compliance levels if at all possible.

I can see that some people in the WG want it, and, if its an API matter, 
not a network matter, that is acceptable.

Getting community feedback is a good plan.

	Andy

[*] because I don't want a partial solution affecting web protocols and 
then requiring backwards compatibility in future recommendations.

Jim Hendler wrote:
> 
> At 11:53 -0500 8/2/04, Dan Connolly wrote:
> 
>> On Mon, 2004-08-02 at 11:40, Thompson, Bryan B. wrote:
>>
>>>  Dan,
>>>
>>>  If it comes to a vote on this, I am going to vote along the lines of
>>>  Jim Hendler and AndyS.  I don't think that we need to added complexity
>>>  in the spec at this time.  So you have my position on this.
>>
>>
>> Umm... well, I can't tell what edits JimH and/or AndyS are proposing.
>> Sorry if I'm reading too fast or something.
>>
>> -- 
>> Dan Connolly, W3C http://www.w3.org/People/Connolly/
> 
> 
> Dan - I won't speak for anyone else, if forced to go with something 
> right now, my proposal would be to include objective 4.5 but remove 
> 4.5.1.  However, I like Kendall's suggestion of leaving both in the 
> current WD and soliciting feedback during review - if we were moving to 
> LC or such I'd worry more, but for a next WD, this seems like an 
> appropriate action to take.
>  -JH
> 
>

Received on Monday, 2 August 2004 15:10:49 UTC