W3C home > Mailing lists > Public > public-linked-json@w3.org > February 2013

RE: Problem with auto-generated fragment IDs for graph names

From: Markus Lanthaler <markus.lanthaler@gmx.net>
Date: Mon, 18 Feb 2013 16:49:42 +0100
To: "'Eric Prud'hommeaux'" <eric@w3.org>
Cc: "'RDF WG'" <public-rdf-wg@w3.org>, "'Linked JSON'" <public-linked-json@w3.org>
Message-ID: <00ef01ce0def$929f2090$b7dd61b0$@lanthaler@gmx.net>
On  Monday, February 18, 2013 2:55 PM, Eric Prud'hommeaux wrote:
> Why would our integrity constraints be different for graph labels as
> opposed to a garden-variety node labels?
>   _:post1 {
>       x:assertions x:expressedAs x:triples .
>   }
> Just as with bnodes in graphs, your system is responsible for
> preserving the fact that the thing which had a sioc:topic "RDF is
> easy" also has some contents which say that RDF statements are
> triples. We can access this information without knowing how the
> system keeps track of this:
>   SELECT ?expression
>    WHERE {
>        ?p a sioc:Post .
>        GRAPH ?p {
>            x:assertions x:expressedAs ?expression
>        }
>    }
> The system can also save this info in a format which other systems
> can consume:
>   { _:0x1234 a sioc:Post .
>     _:0x1234 sioc:topic "RDF is easy" .
>     _:0x1234 sioc:reply _:0xabcd .
>     _:0xabcd a sioc:Post" .
>     _:0xabcd sioc:topic "Named graphs are easy" . }
>   _:0x1234 {
>     x:assertions x:expressedAs x:triples .
>   }

I was talking about the fact that bNodes MUST denote the graph to make this
work. The point is, that if I understand the current semantics correctly,
your example above doesn't mean anything if you would use IRIs because they
do not denote the graph. So, in the default graph you would assert that some
IRI is about a sioc:Post but that doesn't apply to the graph because the IRI
used as a label doesn't denote the graph. So, still assuming I understand
the semantics correctly, currently you can't make any statements about a
named graph. Since Manu's use case requires that, we must specify that
bNodes do denote the graph they label. Otherwise you would associate a
signature with some unknown resource, which may or may not be the graph you
normalized to create the signature.

Could someone please give me a concrete example where it wouldn't make sense
that IRI labeling a named graph denotes it. I'm still struggling to
understand why you would ever need that.

Markus Lanthaler
Received on Monday, 18 February 2013 15:50:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:18:36 UTC