Re: does DAWG actually have time to do WSDL?

Bijan Parsia wrote:
> Ok, I lied.
> 
> Here's a response now :)
> 
> On Mar 21, 2005, at 3:29 PM, Seaborne, Andy wrote:
> 
>> Bijan Parsia wrote:
>>
>>> On Mar 21, 2005, at 12:05 PM, Dan Connolly wrote:
> 
> [snip]
> 
>>>> "For protocol, best guess is maybe 4 Apr for LC candidate."
>>>
>>> Since this strikes me as an implausible date for the query language LC,
>>
>>
>> There should be a draft for the working group by March 31.
> 
> 
> Ok.
> 
>> You can see it the current state at:
>>
>> http://www.w3.org/2001/sw/DataAccess/rq23/
>>
>> Given the size of the doc, there is a limit to how much will change by 
>> March 31!
> 
> 
> I guess it's more a sense from comments I've seen. Significant features 
> and questions of semantics seem under debate. *however*, now that I 
> think about it, all we need to care about is the syntax, for protocol 
> purposes.

One question - is it proposed that the way to send queries over SOAP is the 
XML form or a CDATA string also possible?

> 
>>> and I believe that Protocol is dependent on the query langauge, then 
>>> I find this an implausible date for protocol LC.
>>> Things that need to be completed for protocol (IMHO):
>>>     1) XML syntax for query language with XML Schema description 
>>> (kendall and I are working on that; of course, bit of a moving target 
>>> as the query language keeps changing, or potentially changing)
>>>     2) Sensible XML Schemable XML output format (I thought this was 
>>> the same as the xsi:type discussion, but I'm happy to raise a 
>>> separate issue).
>>
>>
>> Could you pull out what dependences you see that there are on the QL 
>> from the protocol?  I don't see any
> 
> 
> Syntax, primarily. I don't think there's any semantic ones.
> 
>> (I don't put the output format under QL as it's a separate document).
> 
> 
> Ah ok.
> 
> What would make the best WSDL is if we have a plausible XML Scheme 
> element for the input (the query) and the output (the results format). 
> With these, the abstract portion of the WSDL is trivial, as is certainly 
> the SOAP binding.
> 
>> With such a list, I'll make sure that these items are covered soon, if 
>> they are not already done.  The QL is pretty static except for the 
>> syntax and the protocol either takes a string and/or will use an XML 
>> form so that does not seem to be a dependency.
> 
> 
> If the syntax is reasonably static, then yes, we have what we need. Thanks.
> 
>> I'll email something on an abstract syntax ASAP.
> 
> 
> Great.
> 
>> We moved the FROM/WITH out of the QL so the QL has a dependency on 
>> protocol.
> 
> 
> Er...well, not exactly, right? Yes there is functionality that used to 
> be in QL language that now will only be available in the protocol, but 
> moving them out of the query doesn't create a dependancy between any of 
> the remaining aspects of the QL and the protocol, does it? Am I missing 
> something?

Only that it has to support query requests.  There has to be a check that 
the expressivity of that section of the protocol work is sufficient for the 
usage of the QL even if some datasets can't be expressed (that is reasonable 
given that there are many, many configurations of dataset).

Some text to tie to the two together would help the reader.

	Andy

> 
> Cheers,
> Bijan.
>

Received on Monday, 21 March 2005 21:03:37 UTC