W3C home > Mailing lists > Public > semantic-web@w3.org > July 2006

Re: expectations of vocabulary

From: Reto Bachmann-Gmür <reto@gmuer.ch>
Date: Sun, 23 Jul 2006 19:18:42 +0200
Message-ID: <44C3AF72.3040203@gmuer.ch>
To: Henry Story <henry.story@bblfish.net>
CC: Semantic Web <semantic-web@w3.org>
It  seems questionable whether or not a general rule on what triples a
webserver should deliver when an RDF graph is answered makes sense. But
apart from the expectations on the graph the client might have basing on
known triples about the requested resource, the HTTP mechanisms for
content negotiation could be extended for greater expressiveness with
graphs.

A possibility might be an HTTP header indicating the vocabularies
understood by the client, something like:

Accept-Vocabulary: http://www.w3.org/1999/02/22-rdf-syntax-ns#
Accept-Vocabulary: http://www.w3.org/2000/01/rdf-schema#
Accept-Vocabulary: http://xmlns.com/foaf/0.1/

In the HTTP request. The server should then deliver only the triples
with a predicate defined in an accepted vocabulary (similarly for the
class in type-statements). The server can of course deliver inferred
triples to deliver the highest possible amount of information such as by
using a super-property understood by the client.

reto


Henry Story wrote:
>
> RDF solves the mime type explosion problem. Everything can be
> expressed in one of the RDF mime types (I'll use N3 here cause it's
> easier to write). But has the mime type explosion problem just been
> shifted?
>
> Below is an example that should help illustrate the problem.
>
> Example
> -------
>
> Imagine I translate the following atom
>
> <entry>
>    ...
>    <link rel="categories" href="/cats"/>
>    ...
> </entry>
>
> into the following N3
>
> [] a :Entry;
>   iana:categories </cats> .
>
> Where I have defined
>
> @prefix iana: <http://www.iana.org/assignments/relation/> .
> @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
> @prefix owl: <http://www.w3.org/2002/07/owl#> .
>
> iana:categories a owl:DatatypePropery;
>                 rdfs:domain :FeedOrEntry;
>                 rdfs:range :CategoryList .
>
>
> The N3 representation of </cats> could be something like
>
> <> a :CategoryList;
>    :category [ :scheme <http://eg.com/cats/>;
>                :term "dog" ];
>    :category [ :scheme <http://eg.com/cats/>;
>                :term "house" ].
>
>
>
>
> Question
> --------
>
> What guarantee do I have that  the representation returned is not
>
> <> a :McDonaldCategoryList;
>    :McCategory [ :McScheme <http://eg.com/cats/>;
>                :McTerm "dog" ];
>    :McCategory [ :McScheme <http://eg.com/cats/>;
>                :McTerm "house" ].
>
> where the above terms are all defined somewhere as being owl:sameAs
> the ones I was expecting to receive?
>
>
> Solution?
> --------
>
> Well perhaps one could argue that :categories is a relation that
> creates an expectation of things being represented in a certain way.
> But since we are dealing with semantics, that seems a little dodgy.
>
> Perhaps it is just expectations and conventions that we use the same
> vocabulary that will solve this problem. After all that is just what
> happens in normal language learning
> (See perhaps Devid Lewis's book Convention)
>
> Henry
>
>
> Home page: http://bblfish.net/
> Sun Blog: http://blogs.sun.com/bblfish/
>
>
>
>



Received on Sunday, 23 July 2006 17:19:01 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:11 GMT