W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2001

RE: Linking RDF

From: Lee Jonas <lee.jonas@cakehouse.co.uk>
Date: Fri, 20 Apr 2001 15:26:00 +0100
Message-ID: <51ED29F31E20D411AAFD00105A4CD7A770FB@zingiber.cakehouse.co.uk>
To: "'Murray Altheim'" <altheim@eng.sun.com>, Aaron Swartz <aswartz@swartzfam.com>
Cc: Danny Ayers <danny@panlanka.net>, RDF Interest <www-rdf-interest@w3.org>
Murray Altheim [mailto:altheim@eng.sun.com] wrote:

[snip]
>I feel like I'm not making any progress here though in one idea, and 
>that is that linking to "RDF" is almost like linking to "XML" -- one 
>needs to specify what specific grammar of RDF is being served. Otherwise,
>the processor has no understanding of the semantics of the received RDF.
>

I disagree.  RDF is a specific (albeit open-ended, unfortunately) grammar of
XML - any RDF processor can process it.  Making sense of the RDF statements
found therein is another kettle of fish.  RDF 'flavours' (or 'schemas') like
Dublin Core are based upon the namespaces the elements and attributes come
from.  As such, an RDF application can pick out what it knows about and skip
over the rest.

The danger of only supporting specific 'flavours' of RDF such as Dublin
Core, is that you are not really making the specification of metadata within
XHTML open.  You are replacing a closed system with a slightly larger closed
system.  Do you really want to embed knowledge of specific RDF 'flavours'
within separate XHTML modules?  What about mixing and matching Dublin Core
with DAML+OIL metadata?  Taken to an extreme, does this mean a separate
XHTML Module for every permutation of mixed metadata schemes?

>I don't think there's a MIME type for Dublin Core, and of course there
>won't be for author-designed RDF types. So 'type' doesn't really work
>that well either. Ideally, a namespace URI (ironically) would probably
>be best, since that allows for author-designed RDF applications, and
>doesn't force anyone to use only those RDF applications approved by a
>specific body. I advocated XML notations because the NOTATION 
>declarations allow for both FPIs and URIs in declaring 'type'. There's
>currently no attribute on <link> that allows for this type of feature,
>which is why I've never travelled down this road before. But there
>*could* be in XHTML 2.0 if people made enough noise, and convinced the
>HTML WG that this was a good and necessary feature.
>

You don't need a Dublin Core mime-type, it is more appropriate to have an
RDF mime-type, and allow apps to glean info from the different RDF
'flavours' contained therein.  Hence, using a mime-type of 'text/rdf+xml' or
'application/rdf+xml' should do just fine.

>Murray
>
>...........................................................................
>Murray Altheim, SGML/XML Grease Monkey     <mailto:altheim&#64;eng.sun.com>
>XML Technology Center
>Sun Microsystems, 1601 Willow Rd., MS UMPK17-102, Menlo Park, CA 94025
>
>      the wood louse sits on a splinter and sings to the rising sap
>      ain't it awful how winter lingers in springtimes lap -- archy
>

regards

Lee
Received on Friday, 20 April 2001 10:25:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:49 GMT