W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > September 2009

Re: Agenda Topic / Issue: Clarify the meaning of "ignore" with respect to attributes that have no legal value

From: Mark Birbeck <mark.birbeck@webbackplane.com>
Date: Fri, 11 Sep 2009 09:23:33 +0100
Message-ID: <640dd5060909110123x3e010ab4ycb3a1881eac3c042@mail.gmail.com>
To: Tore Eriksson <tore.eriksson@po.rd.taisho.co.jp>
Cc: public-rdf-in-xhtml-tf@w3.org, Shane McCarron <shane@aptest.com>
Hi Tore,

> My understanding from reading "Named Graphs, Provenance and Trust" [1],
> is that blank nodes can't be shared between graphs, making this type of
> query impossible. Graph A has to be a superset of the default graph for
> this to work.
> [1] http://www2005.org/cdrom/docs/p613.pdf

Yes, that's true.

I think not being able to share bnodes across named graphs retrieved
from arbitrary locations is a reasonable limitation, since the very
definition of a bnode implies that they are local to a particular

However, the situation I am describing is one in which all triples
originate from the same graph -- the document currently being parsed.
This means that all bnodes are known to originate from the same
document, and so can be shared amongst any graphs that the RDFa parser
chooses to create.

The only reason for the second graph is to keep parser-specific
additions out of the default graph, for conformance purposes.

Another way we could look at this would be to allow a graph to contain
further graphs; we parse document A, to give us named graph A, and
this graph in turn comprises further graphs -- the default graph, as
required by the RDFa spec, and further graphs as deemed appropriate by
the parser.

>From a SPARQL point of view, to do what I described before (get all
triples relating to document A) you'd first query for the names of all
graphs that relate to graph A, and then run a query against all of
those graphs at the same time.

We'd probably need a flag in the query to tell SPARQL whether or not
to treat bnodes as being the same across different graphs, but I guess
that's the kind of thing that would get discussed if we move to trying
to formalise named graphs.



Mark Birbeck, webBackplane



webBackplane is a trading name of Backplane Ltd. (company number
05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
London, EC2A 4RR)
Received on Friday, 11 September 2009 08:24:16 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:02:04 UTC