W3C home > Mailing lists > Public > www-rdf-interest@w3.org > March 2003

Re: Using MIME types in RDF/DAML and Image Metadata

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Mon, 31 Mar 2003 11:39:09 +0100
To: Yarden Katz <katz@underlevel.net>
cc: www-rdf-interest@w3.org
Message-ID: <3211.1049107149@hoth.ilrt.bris.ac.uk>

>>>Yarden Katz said:
> Dave Beckett said:
>>   Mapping between URIs and Internet Media Types
>>   TAG Finding 8 April 2002 (Revised 27 May 2002)
>>   http://www.w3.org/2001/tag/2002/01-uriMediaType-9
> As I pointed out in my original post:
>  Googling showed up a seemingly relevant link
>  "Mapping Between Content-Types (MIME types) and URIs"
> The W3C document you made reference to contains two broken links to
> IETF documents, one of which I mentioned above.

Yes, those IETF internet draft URIs are never "Cool URIs".

But the date of the TAG finding - May 2002 - is much newer than the
2001 I-D and won't be going away.  But you've got a solution:

> > The other way uses familiar and well deployed Dublin Core vocabulary
> > an it's dc:format property.  It takes as primary recommend value the
> > IMT / MIME Type as a literal.  You might be able to do some inference
> > magic to imply that all your 'Picture's have the dc:format
> > "image/jpeg".  Otherwise you'd need to put it in the instance data.
> Great!  I never knew about this property of DC.  However, I'm not
> quite sure I understand your suggestion about inferencing.  As far as
> I can tell, I could use dc:format as follows:
>   Given: <Picture rdf:resource="http://foo/bar.jpg">
>   <rdf:Description rdf:resource="http://foo/bar.jpg">
>     <dc:format>image/jpeg</dc:format>
>   </rdf:Description>

That's right

> But I'm unsure of a way to express that "all resources pointed to by
> [i.e., everything in the rdfs:range] 'Picture' have dc:format of
> image/jpeg".  If I hardcode it into the agent, well, then it's not
> really inferencing :)

I thought since you were using daml:ObjectProperty that you had a
reason to use it, and understood it (I don't much).  If that isn't
the case, I'd say stick with rdf:Property.  If I understand it right,
DAML+OIL or OWL could let you make this level of inferencing.


Received on Monday, 31 March 2003 05:39:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:41 UTC