W3C home > Mailing lists > Public > public-rdf-wg@w3.org > March 2013

Re: scope review and an altenative phrasing.

From: Pat Hayes <phayes@ihmc.us>
Date: Sun, 17 Mar 2013 19:42:48 -0500
Cc: Antoine Zimmermann <antoine.zimmermann@emse.fr>, RDF WG <public-rdf-wg@w3.org>
Message-Id: <3B3BDBE0-7268-4146-BC74-356A1AA11939@ihmc.us>
To: "Peter F. Patel-Schneider" <pfpschneider@gmail.com>

On Mar 17, 2013, at 3:49 PM, Peter F. Patel-Schneider wrote:

> Somehow I'm missing the reason that scope has to show up in Semantics at all.

I have come to think you are right. But I think *something* needs to be defined (see below) and that it should probably be in Concepts. 

> Isn't it sufficient to just say the bnode mapping in the semantics is from the bnodes in the graph and that combining RDF graphs takes their union?  If the graphs don't share bnodes then you get the standard behaviour.  If the graphs to share bnodes then you get the bnode-sharing behaviour.


>  The condition on the lemma stating that a set of graphs entails their union then is expressed in terms of the union, which turns out to be something like for each bnode in the union one of the constituent graphs contains all the triples in the union for that bnode.  No need to talk bring forward any other machinery.

Yes, that works. A closely similar alternative is the one I described in a reply to Ivan, which is to define the notion of a complete graph (for every blank node, if it contains a blank node then it contains all triples which contain the blank node) and observe that any set of complete graphs entails and is entailed by its union. 

Your way might be simpler, though, I have to admit. And it is actually more general. OK, lets go with that. I will put out a new draft later tonight using this. 

> This, of course, punts on how bnodes are identified in surface syntaxes.  That's up to the particular surface syntax, of course, but Concepts can provide guidance, and bring forward the notion of the scope of bnode identifiers.

Indeed. But there is still this old issue of, how do we state that two unrelated graphs don't share a blank node by accident, as it were? One way is to require that each complete surface-syntax document descibes a complete graph. 


> peter

IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Monday, 18 March 2013 00:43:27 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:04:26 UTC