W3C home > Mailing lists > Public > www-rdf-interest@w3.org > December 2000

Re: Putting context in RDF serialization

From: Seth Russell <seth@robustai.net>
Date: Wed, 20 Dec 2000 13:55:45 -0800
Message-ID: <3A412AE1.281CFF9D@robustai.net>
To: "Sean B. Palmer" <sean@mysterylights.com>
CC: RDF-IG <www-rdf-interest@w3.org>
"Sean B. Palmer" wrote:

> Good idea. In the context Schema, we could define further context models
> for whatever assertions we have. For example, in your model:-
>
> [contextUri3]--subContextOf-->[contextUri1]
> [contextUri3]--subContextOf-->[contextUri2]
> [contextUri3]--contextFor-->[id1,http:..../Lassila, Creator, "Ora Lassila"]
>
> This is a Schema itself, but one that points "backwards" from the actual
> context Schemas.

Just for the sake of clarity, I think I'd better define what I mean by schema.
I've been calling everything that hangs off the property nodes "schema".   In
other words in :

[s1, p1, o1]

[p1, p2, o2]
....

Everything associated by the same p? subject node is (to me) a schema.
Hopefully I haven't been playing too fast a free with the term.    If you buy
that, then you would need to admit that my context nodes are not schema
nodes.   There would, of course, be two schema nodes that applied to context:

[contextFor, p?, o?] ...
and
[subContextOf, p?, o?] ...

> As such, contextUri3 could itself be the context of
> another context Schema:-
>
> [contextUri4]--subContextOf-->[contextUri3]
>
> In that case would it automatically import the "contextFor" triple that is
> included in contextUri3 for the purposes of the RDF code you made?

Yes that was the idea.

> In other
> words, would this example have the same context as yours(?):-
>
> <?xml version="1.0"?>
>   <RDF
>     xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>     xmlns:s="http://description.org/schema/"
>     contextFor="contextUri4"
>    >
>     <Description about="http://www.w3.org/Home/Lassila">
>       <s:Creator>Ora Lassila</s:Creator>
>     </Description>
>   </RDF>
>
> If so, then it's lucky you use contextFor in the literal sense rather than
> just pointing out a node!

Huh?  I meant "contextUri(n)" to stand for some URI that points out a node.

> In summary, if these are your proposals, then I think that they are very
> useful indeed and should probably be added as RDF serialization properties.

To be honest, most of this I got from Graham.  But Graham didn't specify how it
was to be serialized.  If we have to specify the context of information by RDF
bags and reified statements, then I think it will be just too tedious for
practical use .. so why not just imply all the details by a very simple easily
understood syntax.

Thanks for the dialogue ...
Seth Russell
Received on Wednesday, 20 December 2000 16:51:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:47 GMT