W3C home > Mailing lists > Public > public-sparql-dev@w3.org > July to September 2010

Re: First order logic and SPARQL

From: Axel Polleres <axel.polleres@deri.org>
Date: Sun, 5 Sep 2010 16:05:59 +0100
Cc: "Bob MacGregor" <bob.macgregor@gmail.com>, "Juan Sequeda" <juanfederico@gmail.com>, "Jitao Yang" <jitao.yang@gmail.com>, <public-sparql-dev@w3.org>
Message-Id: <B91B0FF2-D977-49A5-9075-4E2A6A953167@deri.org>
To: Sandro Hawke <sandro@w3.org>

On 5 Sep 2010, at 05:09, Sandro Hawke wrote:

> On Sat, 2010-09-04 at 18:29 -0700, Bob MacGregor wrote:
> > It would necessarily
> > have closed-world semantics (as does Datalog).
> 
> My understanding is that in datalog you can't tell if it's closed or
> open world, since there's no negation (or negation-like operators, such
> as OPTIONAL).   As such, datalog can be a subset of both prolog and FOL.
> 
> Am I wrong about that?

Well yes... extensions of Datalog by negation as failure are quite standard.
Particularly, it shoudl be quite common knowledge in between that this is 
needed to capture SPARQL's OPTIONAL: These papers [1,2] show that SPARQL can be 
captured by Datalog with stratified negation as failure and in fact (a slightly 
simplified version of - i.e. not bag semantics, AFAIR) SPARQL has the same expressive 
power of Datalog with negation as failure.

HTH,
Axel

1. Renzo Angles, Claudio Gutierrez: The Expressive Power of SPARQL. International Semantic Web Conference 2008: 114-129

2. Axel Polleres: From SPARQL to rules (and back). WWW 2007: 787-796


> 
> Meanwhile, I'm listening to the discussion about SPARQL semantics, but
> don't have any to add right now.
> 
> [Note that list policy is try to reduce cross-posting [1].  I'm dropping
> semantic-web@w3.org, because I think this is mostly about SPARQL.]
> 
>    -- Sandro
> 
> [1] http://www.w3.org/Mail/
> 
> 
> 
Received on Sunday, 5 September 2010 15:06:34 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:15:50 UTC