I have been told of a W3C Note written by Mark Birbeck of x-port.net, that discusses linkage between XHTML and RDF. Published on 14 February 2004, it seems to address our issues very well.
The ideas it proposes come from current work on XHTML 2.0 which is the current focus of work for the primary Mark Up language of the near-future. (we’re now on XHTML 1.1). A feature of XHTML is greater support for technologies like RDF.
The W3C Note speaks for itself but if you’ll allow a little interpretation from me, seems to boil down to the following possible expressions of content ratings etc. for us.
As a link tag
In other words, all the discussion in the document that came out of the first meeting [Tech1] can be reduced to adding the about attribute to the link tag. The client receiving this still has to understand that the about attribute is to be taken as such for the RDF found at the given href but this seems easier and clearer than adding the same information as a query string after the URL.
Anything written in a link tag can be written as an HTTP Response Header so the equivalent of the above would be:
Link: ; /=”/”; rel=”meta”; about=””; type=””application/rdf+xml”
Previous discussions about creating link tags that point to specific ratings or to a file that contains both the ratings and the rules for their application still apply.
As meta tags
Birbeck concentrates on writing RDF metadata not in XML but in N3 notation and within meta tags. Since, unlike its predecessors, XHTML2.0 allows meta tags to be nested, we get an example like this:
A page with embedded ratings 1 1 1 1 1
(The “rating” structure is probably not necessary but I’ve kept it in for consistency with other examples).
So if we want to embed RDF within XHTML, this seems to be the way to do it – but as noted in Tech1, it would mean supporting two methods for encoding the ratings in any filtering client, as well as presenting webmasters with a choice many won’t understand.
It is, however, more robust than the simple and current practice of, for example:
The new method allows reference to be made to external resources. So a webmaster might have something like this:
John Smith
So the potential here is for a lot of flexibility and different solutions will suit different webmasters.
Labelling other things
Birbeck raises an intriguing possibility – that of adding a property value to any HTML tag. For our purposes, this might make the following possible:
- label a hyperlink (a filter would then make it impossible to follow that link if the rating showed it to be inappropriate for the user),
- label images directly from the XHTML that calls them rather than through an HTTP Response header (different circumstances will suit different people)
- Label banner ads again at the point at which they’re called by the containing page.
This is rather speculative but the job of the working group is to speculate!
References
W3C Note http://www.w3.org/MarkUp/2004/02/xhtml-rdf.html
Tech1 http://www.icra.org/press/labellingWG/tech1/