W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2010

Re: Another thing on property paths...

From: Axel Polleres <axel.polleres@deri.org>
Date: Fri, 17 Dec 2010 14:17:03 +0000
Cc: "SPARQL Working Group" <public-rdf-dawg@w3.org>
Message-Id: <C0920BAC-D741-43FB-958E-841A0BD936E5@deri.org>
To: "Andy Seaborne" <andy.seaborne@epimorphics.com>

On 17 Dec 2010, at 13:23, Andy Seaborne wrote:

> 
> 
> On 17/12/10 12:14, Axel Polleres wrote:
> > On 17 Dec 2010, at 10:54, Andy Seaborne wrote:
> >> What changes to
> >>
> >> http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath
> >>
> >> are you proposing?
> >
> > Do I understand correctly, that the current definition does exactly what I propose? :-)
> 
> I am asking you what changes would be necessary.

> We need to work in a mode where all comments/issues are related to the
> document.  The starting point is the text that is currently there.

I surely agree, let me try again...

1)
I wrote what I proposed to change immediately after that line in my mail.
For your convenience, let me try here again slightly reformulated:

I am *not* talking about the definition 
 http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath
which seems to do exactly what I fined intuitive, I am actually talking 
about the definition in in Section 18.1.6.:
http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#sparqlPropertyPaths 

"
Definition: ZeroLengthPath

A zero length path matches all subjects and all objects in the graph, 
           and also any IRIs explictly given as endpoints of the path pattern.
"
This sounded to me like Cartesian product as opposed to the formal algebra definition you cite.
Can I suggest to reformulate this as:

"
Definition: ZeroLengthPath

A zero length path matches all subjects, objects in the graph, and also any IRIs explictly given as endpoints of the path pattern,
to themselves."

2) 
I *do* have some open question with regards to 
  http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#defn_evalZeroPath
regarding nodes(G).

 Since I didn't find a definition of nodes(G), do you imply only the terms used as subjects/objects in G there or also the terms used as predicates?

3) 
> > BTW, I am not clear why you include IRIs mentioned in SERVICE patterns, what about IRIs explicitly mentioned in GRAPH
> > patterns then? Shouldn't there at least be aremark about why those aren't matched
> 
> Where?
> 
> "endpoint of the path" seems clear enough to me.  Nothing to do with 
> SERVICE.


well, admittedly, it wasn't, i.e. isn't "and also any IRIs explictly given as endpoints of the path pattern." just redundant then with
"all subjects and all objects in the graph" ?!?


Axel
Received on Friday, 17 December 2010 14:17:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:44 GMT