Outcome of Meeting with Jeremy Carroll, Stuart Williams and Dan Brickley

Hi all,

(SWCG as input to telecon on Friday).

As previously discussed, JJC, Dan Bri and Stuart W kindly gave up 
practically their whole day to look at the POWDER issues yesterday (and 
Jeremy's been posting a lot of stuff to www-archive recently too [1]).

First of all I would like to record my thanks to them and to HP for 
their hospitality.

I've tried to capture the output of that meeting as best I can in a 
document that sets out the problem (as I understand them) and the 
solution that is now being proposed. It would be wrong to post that doc 
on the group's web space (I've been in trouble for doing something 
similar before :-) ) so I've posted it at [2] and included some 
disclaimers. Any and all errors are mine.

In brief - we do without reification and named graphs, we avoid non-SW 
specialists having to handle complex OWL constructs and we use GRDDL to 
anchor simplified RDF/XML documents that have a limited semantics in 
something that is consistent with formal logic. (Note to self: take 
carrot when next visiting Bristol to feed the rabbits that can sometimes 
appear from underneath Jeremy's hat).

We noted some action items which I'd like to record here (and they'll 
make sense in the context of [2]):

Action: Dan to write warning paragraph concerning risks of creating 
blank nodes with properties.

Action: JJC to write output, i.e. verbose RDF/OWL doc that says the same 
as the example.

WG: Include notes on the declarative semantics of pre and post transform 
in the documents. 'Operational semantics' applies to pre-transform (what 
in [2] I call POWDER Lite), 'formal semantics' apply only to 
post-transform (POWDER Full).

POWDER Full allows more sophisticated reasoning, profile matching. 
Importantly, it also provides the academic rigour for the pre-transform 
version (POWDER Lite).

Action: JJC to write to POWDER group to explain the problem of dealing 
with sets of resources cf. sets of URIs. (end result is likely to be 
POWDER only applies to sets of URIs).

Action: JJC to re-draft formal semantics [3] to match URI-based approach.



[1] http://lists.w3.org/Archives/Public/www-archive/2007Dec/
[2] http://www.fosi.org/projects/powder/bristol.html
[3] http://lists.w3.org/Archives/Public/www-archive/2007Dec/0042.html



-- 
Phil Archer
Chief Technical Officer,
Family Online Safety Institute
w. http://www.fosi.org/people/philarcher/

Received on Wednesday, 19 December 2007 13:29:42 UTC