- From: Graham Klyne <GK@NineByNine.org>
- Date: Thu, 06 Jun 2002 12:31:33 +0100
- To: "Butler, Mark" <Mark_Butler@hplb.hpl.hp.com>
- Cc: "'Art.Barstow@nokia.com'" <Art.Barstow@nokia.com>, w3c-ccpp-wg@w3.org, www-mobile@w3.org
At 09:53 AM 6/6/02 +0100, Butler, Mark wrote: > > Why don't you guys just re-do CC/PP in RELAX-NG or XML Schema :-) > >"Why don't you guys just" ... if only it was that easy :-) > >Two problems >i) The CC/PP charter says CC/PP has to be based on RDF. >ii) The CC/PP charter says it has to be backward compatible with UAProf. > >So we can't make changes like this without changing the charter and getting >agreement from WAP Forum. (We also had this debate early in the life of CC/PP.) If CC/PP simply wanted to go away and play in its own sandbox, without regard for other developments happening around the Web, then that would be a plausible approach. From my perspective, RDF is the first attempt to explicitly recognize that many of the protocols that are developed in isolation really need to work together, and that end is promoted by having some common framework. XML is a start, but it operates only at the syntactic level: common parsers can be used but the application logic is still anything but common. We have already seen debates about how CC/PP should interact with P3P. I understand that P3P was originally intended to be RDF-based, but the WG backed away from that decision for reasons similar to those being aired now for CC/PP. But that was, in my view, a short-sighted decision which came home to roost when we came to look at CC/PP and P3P interactions: if P3P had been built as a true RDF application then many of the difficulties noted would not have arisen. (Since then, work has been done on P3P to map it into RDF, which is a step in the right direction, but there is still a translation step involved which still leaves the barrier to integration higher than it would be if these "metadata protocols" were consistently developed in RDF - or some other common framework - which allows sharing of application logic as well as syntax.) I'll also note that CC/PP isn't perfect in this regard: e.g. many of the attribute values inherited from UAProf would, in my hindsighted opinion, be better represented using RDF primitives. But, hey, we had to start somewhere! And there was a charter constraint in that regard. #g ------------------- Graham Klyne <GK@NineByNine.org>
Received on Thursday, 6 June 2002 07:18:50 UTC