- From: Butler, Mark <Mark_Butler@hplb.hpl.hp.com>
- Date: Fri, 17 Aug 2001 11:39:14 +0100
- To: www-mobile@w3.org
Of course the big advantage of CC/PP is that schemes like Morphis & Cocoon do not provide "plug and play" i.e. they cannot adapt content for devices which are not in their device database. CC/PP allows device manufacturers to provide information about new devices. That's why it would be better if programs like Morphis or Cocoon used CC/PP rather than proprietary XML formats. The issue about whether to use device classes or not (how to choose your axis of negotiation) is very important and central in my view to vocabulary design. The more axis of negotiation, the more flexibility you have (to paraphrase Johan the more edge cases you can deal with) but also the more work you need to do provide adaptations of pages or alternate versions of content on pages. My guess is simple device classes is too little but every attribute of the device will be more than the content author / servlet author can be bothered with so you need something in between. IMHO I think someone (maybe the CC/PP WG, maybe the Device Independent WG) needs to think about issues like this and design some standardised vocabularies which explicitly describe how attributes can be used to adapt content for target devices. I think there is a danger that if you don't think about the adaption issue, you can run into problems later. For example in UAProf there is an attribute called WmlDeckSize i.e. the max size of a deck (the WML equivalent of a HTML page) that a phone will accept. However the manufacturers have not agreed whether this value refers to the WML file, a compressed WML file, or even a WML file with a reply header. Therefore if someone writes a transcoding proxy in order to turn HTML pages into WML and chunk them up where necessary into smaller pages they cannot rely on WmlDeckSize to tell them the maximum size of page to output. A few WAP Forum members have noticed this problem so hopefully they will clarify the standard before it is too late. Mark Butler -----Original Message----- From: Johan Hjelm [mailto:johan.hjelm@era-t.ericsson.se] Sent: 17 August 2001 01:33 To: olivier.rovellotti@3tl.com Cc: 'Butler, Mark'; 'cafe babe'; www-mobile@w3.org Subject: Re: Any CC/PP App servers in the market Actually, they create a device class... Same problem as before. It is always the edge cases that get you. Olivier Rovellotti wrote: > Mark, > > Thanks for the pointers, they both are very interesting approch, and > specially the morphis effort. > The information provided by Argogroup is good althought very incomplete. > > The XML profiles solution of (http://www.morphis.org/) have a lot more to > offer so far. > I am sorry to say that I would be more keen to parse their XML descriptions > of a few hundred devices then one or two UA-Prof description :-( > > It seems that for the CC/PP-UAProf to be succefull is would really reuired > more profiles to be available > > Sorry to state the obvious >
Received on Friday, 17 August 2001 06:39:19 UTC