Device independence and RDF query

I was reading Chris Lilly's comments [1] w.r.t. a TAG position on new TLDs; 
his phrasing suggested an approach to one of the problems that, AFAIK, has 
not really been addressed with respect to the CC/PP and device independence 
work, namely how to use a CC/PP profile to drive content selection and 
adaptation.

I'm not sure where is the best place to raise this and don't want to 
cross-post all over, so I'm hoping that my mentioning it here it may filter 
to interested parties.   One one hand, it might be seen as a use-case for 
the query component of DAWG, or it might be input to the device 
independence work, or it might be seen as a functional requirement on 
future stylesheet work (with a view to more effectively combining XML and 
RDF views of data).  There's been some recent discussion of using XSLT to 
bridge from XML to RDF;  this suggestions suggests that XSLT might also be 
able to work the other way.

At 05:09 11/05/04 +0200, Chris Lilley wrote:
>I strongly suggest adding, either here or right after the paragraph
>about CC/PP,
>
>   Use of the CSS @media construct allows different styles suitable for
>   desktops (@media screen) and mobile handheld devices (@media
>   handheld). CSS media queries, the client-side complement to
>   server-side CC/PP, gives finer grained stylistic control.
>
>CC/PP allows a client to give a server delivery context information to 
>help with
>server-side styling and content filtering; CSS @media and media
>queries is the necessary client-side component of that where the
>content author can specify a range of stylings appropriate to
>different devices. These two approaches can of course be used
>together. The client-side approach is sometimes the only one possible,
>for example in p2p or messaging applications when there is no server,
>or the server is a micro-server on a phone and not capable of content
>transformation and filtering.

Chris uses the term "media queries", and it occurs to me that a way to 
drive content selection might be for a stylesheet processor to incorporate 
an RDF query processor that can be used to query the presented CC/PP client 
profile, possibly with variable bindings from the query being fed into the 
subsequent styling operations.

This suggests a stylesheet processing model that looks something like this:

    +-----------+     (Xpath-based selection)
    | XML input |->-
    +-----------+    \     +--------+
                     (*)->-| Output |
    +-----------+    /     +--------+
    | RDF input |->-
    +-----------+     (RDF query based selection)

Work on making RDF queries fit in the XPath framework might be deployed to 
make this work more uniformly.

The CC/PP use case that suggests this approach would probably require the 
RDF queries to operate over an inferred RDF graph (e.g. to resolve CC/PP 
default values), so a generic processing tool along these lines would 
incorporate (say) XSLT, RDF query and RDF rules.  As a first approximation, 
imagine, say, a version of CWM integrated with an XSLT processor.

Simple RDF query processing is not hard to do, so this might prove to be a 
simple and flexible approach to providing a standard way to mesh content 
adaptation with client capability descriptions.  Inference is a little 
harder, but can be reasonably easily to express in a functional language 
(which XSLT is) given the right query primitives, so this doesn't look like 
a vast leap beyond currently exists;  I think most of the details designs 
already exist.

#g
--

[1] http://lists.w3.org/Archives/Public/www-tag/2004May/0023.html

PS:  Another approach to content selection (more akin to my work in IETF 
RFC 2533) might be to adapt an OWL/Description Logic approach, using 
subsumption as the mechanism for matching client descriptions with content 
descriptions, but that doesn't sit easily with the current CC/PP format.


------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact

Received on Tuesday, 11 May 2004 06:21:32 UTC