At 08:53 PM 1/10/03 +0000, Phil Archer wrote: >This is certainly very interesting - you seem to be addressing some of the >problems a little group I'm coordinating is looking at too, namely how to >lead a client to an RDF description that can then be applied to everything >else on the site. HTTP's stateless nature means that, of course, images >within a web page wouldn't be covered by a description that is *about* the >page and so on. This is a show-stopper as far as using RDF for rating and >filtering in the area of child-protection is concerned. P3P has some mechanisms of this ilk [1]. >We're only just getting going but the ideas in my mind are that an rdf:about >element might be able to include wildcards and regular expressions. Ouch! I would anticipate, rather, that a rule language might find a way to express axiom schema. >Another idea I want to explore (which may be unworkable but I want to know >it's unworkable before I give it up) is that, say, a film classification >board could lodge an RDF description of its various categories. Streaming >media could then point to one of those limited number of classifications >which in turn would recognise the pointer as valid (through some sort of >DSig no doubt). The descriptions would be cached just about everywhere >before long. Sounds a bit like CC/PP device profiles doesn't it? Hmmm... in the IETF conneg work, we came up with a device for identifying feature sets using a name based on a hash of the corresponding feature set [2] (based on [3]). This gives potential for self-validating caches of feature descriptions. #g -- [1] http://www.w3.org/TR/P3P/ [2] http://www.ietf.org/rfc/rfc2938.txt [3] http://www.ietf.org/rfc/rfc2533.txt ------------------- Graham Klyne <GK@NineByNine.org>Received on Monday, 13 January 2003 08:34:59 UTC
This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:44:40 UTC