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

Re: Go ahead with pub

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Tue, 3 Oct 2006 15:33:55 +0100
Message-Id: <FFFCB4F8-C0D6-4355-A22F-92A4E7D6C99F@cs.man.ac.uk>
Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>
To: Eric Prud'hommeaux <eric@w3.org>

On Oct 3, 2006, at 3:11 PM, Eric Prud'hommeaux wrote:

> On Mon, Oct 02, 2006 at 10:14:14AM +0100, Bijan Parsia wrote:
>> I think the status part should point to the issues list.
>>
>> """Costs: Tableau-based reasoners (at least, the Pellet Demo example
>> 7) rely on the current, more expressive semantics to match
>> implications that are not in a materializable RDF graph."""
>>
>> No. Pellet uses BNodes as syntax for non-distinguished variables, as
>> that's what we were told was the likely syntax for non-distinguished
>> variables in SPARQL/DL. The semantics of *all* variables in SPARQL/
>> RDF is semi-distinguished.
>
> But we need to explain to the community, in simple terms, why we

Need?

> non-distinguished variables.

Not really. Or if we do, this is not the vector, nor is the text  
above a good way to do it.

>> I thought the alternative proposal (e.g., from conversation with
>> Jeen, Jorge and others) was to *drop* BNodes in triple patterns. That
>> does solve all the problems of scope, meaning etc., but it means that
>> certain combinations of the axes of distinguishedness will be harder
>> to specify (but heck, we can always introduce syntax later).
>
> The majority of the deployed

?

> SPARQL and RDQL

RDQL doesn't support BNodes in patterns, which is what we're talking  
about.

> implementations

Not so sure about that. But anyway, so what?

> follow
> the subgraph semantics of [LC1]

I'm not getting what the semantics have to do with it.

> so I am trying to explain why we have
> a different semantics.

Of what? What does this have to do with the syntactic issues.

> This includes why we were motivated to use the
> E-entailment semantics of [LC2] and test cases [DIF] to show how they
> might differ.

I don't see how this is relevant to the above issue. Jeen and Jorge  
don't want to get rid of BNodes because of DLs!

> Not having BNodes at all seems like a last resort,

I believe most RDF query languages don't have BNodes in patterns  
(I've not checked). RDQL doesn't. So I don't see that it's a pressing  
feature such that the removal is a last resort.

And regardless, I don't see how it helps to mistate things. Simpler  
terms, maybe; *wrong* terms, not at all.

> if we discover that
> a large fraction of the community is specifically dis-served by us
> keeping them in.

I stand by my point that that bit is wrong and thus confusing. That  
you meant to introduce and explain *other* issues by it makes it more  
confusing. So, let's strike it.

Bijan.
Received on Tuesday, 3 October 2006 14:34:59 GMT

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