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

Re: summary of some cwm/euler implementation experience w.r.t. accessing RDF collections

From: Seaborne, Andy <andy.seaborne@hp.com>
Date: Tue, 08 Nov 2005 12:10:10 +0000
Message-ID: <437095A2.3020103@hp.com>
To: Dan Connolly <connolly@w3.org>
CC: jos.deroo@agfa.com, RDF Data Access Working Group <public-rdf-dawg@w3.org>

FYI:

ARQ now implements calculated properties like:

{ [] :values ?someList .
   ?someList :listMember ?item }

which will generate solutions for ?item.

If there is common property defintions, I'll support them.


Same for members of bag/alt/seq s.


Implemention experience:

There is an interaction with queries and inference graph (the logical closure 
of the input graph with respect to some vocabulary) in that the inference 
engine might also be doing this to create triples that also encode that 
information.  There is a chance both will do it :-) resulting in some 
inefficency (and so the feature is off by default).

It's looks like a bad match to plain SQL - stored procedurs or recurisve SQL 
may help.

	Andy


Dan Connolly wrote:
> Jos,
> 
> I polled the WG on this, and there seems to be a critical
> mass of support to do this as a Working Draft.
> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/att-0145/01-dawg-minutes.html#item03
> 
> 
> EricP expreseed interest in helping... I wonder if this would
> compete with getting SPARQL QL stuff done, or if it would provide
> a needed fun distraction?
> 
> Aside from EricP, there weren't many offers to help with the writing,
> but people were supportive of making it a WG deliverable; they would
> either review it or trust that it was OK.
> 
> I'd be particularly greatful if you could elaborate your
> sketch to address this comment about traversing trees...
> 2005-10-26T14:53:28Z from david.h.jones
> http://www.w3.org/mid/01CF21867FABC44EBFAC57024D472BEB0189261D@XCH-NW-2V2.nw.nos.boeing.com
> (and following).
> 
> I think trees can be done with inference in much the same
> way as lists.
> 
> It would be nice if we could request publication as a Working Draft
> before the 21 Nov pre-AC-meeting publication deadline. In order
> to do that, we'd need a WG decision to publish in our 14 Nov meeting,
> so we'd need something to look at by 13 Nov, at the latest.
> It wouldn't have to be *completely* finished, but it would
> have to be close enough that the WG would trust you and
> maybe one or two other people to finish it up.
> 
> 
> On Sun, 2005-10-09 at 09:42 -0500, Dan Connolly wrote:
> 
>>On Sun, 2005-10-09 at 15:40 +0200, jos.deroo@agfa.com wrote:
>>
>>>In message 
>>>http://lists.w3.org/Archives/Public/public-cwm-talk/2005OctDec/0002.html
>>>I promised Dan to spare some time to write a short note about accessing
>>>RDF collections and I found some sunday afternoon time to write
>>>http://lists.w3.org/Archives/Public/www-archive/2005Oct/att-0013/arc.html
>>>which is just an initial attempt :-)
>>
>>Hey, that's a great start.
>>
>>Alistair, can you take a look and see if this is helpful?
>>It's pretty much what I had in mind.
>>
>>We'd just add some title page stuff and references,
>>and publish it.
>>
>>Maybe one or two more examples would be useful.
>>
>>Alistair, can you give us an example from an application
>>you have experience with?
>>
>>I think I have a banking example somewhere, with a list
>>of checks or something. I'll try to find that one.
>>
Received on Tuesday, 8 November 2005 12:11:16 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:24 GMT