W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2006

Re: Agenda request: characterize the diffs between subgraph-matching and E-entailment

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Mon, 9 Oct 2006 21:12:11 +0100
Message-Id: <1BC01AF9-873E-475C-AE62-07A33306BC12@cs.man.ac.uk>
Cc: "Seaborne, Andy" <andy.seaborne@hp.com>, RDF Data Access Working Group <public-rdf-dawg@w3.org>
To: Bijan Parsia <bparsia@isr.umd.edu>

Let me just make my "ordering of effort" request clear.

 From what I can tell, there are issues with both the specification  
of BGP matching and with the algebra (e.g., filter variable scope,  
idempotency, communitivity, paradoxical queries), and in both cases,  
there are both semantic and presentational issues.

It is my organizations belief that semantic issues *must* be fixed,  
and presentational issues *should* be fixed. We are willing to  
jettison features if required. For example, I gave up on my preferred  
notion of distinct, even as an option. I have worries about Graph  
variables (mostly because they have not been adequately examined). Etc.

This week, I've been heads down in the algebra, reading several  
papers, trying to figure stuff out from the spec. I'm behind, alas,  
as I hoped to have something out earlier today. But I'm behind in  
part because of having to deal with that damn pink box and various  
confusions engendered by it.

So please, I would love to see test cases illuminating the  
differences between LC1 and LC2. No problems there. But "counting"?  
Confusing it with how non-distinguished variables act in Pellet?  
Confusing behavior with specification mechanism? That's not helpful,  
IMHO.

I believe section 4 of "The semantics and complexity of SPARQL" and  
Fred's various papers are closely related (i.e., relating issues of  
top down/constuctive vs. bottom up/destructive (I may not have gotten  
the terminology aligned correctly) and how this affects the answers  
especially when dealing with optional). I welcome other eyeballs on  
this.

If the WG would like to turn to the semantic framework right now, we  
can do this, but then I lose a lot of that swapping in. Not to be  
toooooooo snotty, but I'm the most active person who has a strong  
grip on the semantic framework and how it relates to OWL, RDFS, etc.,  
at the moment. So, I shall be even more frustrated if my effort to  
get swapped in on the algebra gets even more diverted.

Cheers,
Bijan.
Received on Monday, 9 October 2006 20:12:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:27 GMT