- From: Charles McCathieNevile <charles@sidar.org>
- Date: Mon, 13 Jan 2003 10:44:53 +1100
- To: Nick Kew <nick@webthing.com>
- Cc: w3c-wai-ig@w3.org
On Monday, Jan 13, 2003, at 10:31 Australia/Melbourne, Nick Kew wrote: > On Mon, 13 Jan 2003, Charles McCathieNevile wrote: >> you should probably look at CC/PP [1] as similar work > > Are you sure about that? My concern is that RDF introduces a whole > new level of complexity, that doesn't really seem justified in > such a simple negotiation. Unless perhaps we reduce it in the > manner of dc: in HTML <meta> elements? > > One of the objectives of mod_accessibility is to be a low-overhead > solution: it is by nature *much* faster and less obtrusive than betsie. > HTTP headers are well-suited to a lean-and-mean system; RDF less so. I am sure you should look at CC/PP... Actually I am not certain that the use of RDF does increase the overhead unjustifiably. The problem with using anything else is keeping track of the possibilities (what is the relationship between a 'betsie' view and a linearised one, ...) and how they relate. The size of the overhead is an important question, but there are limits to what can sensibly be done by using X-some-unknown-extension and on a wide scale it seems to me that something more than plain text strings is important. >> (it is about >> providing profiles of user agents to make sure they get sent >> appropriate content, which I think is pretty much what you are talking >> about). > > Not entirely. The idea is not that the browser has a preferred version > (though it may do within different renditions), but that the user will > wish to switch frequently between different versions. Exactly. CC/PP does allow for this - preferences are not necessarily applied for more than a single request, and so you could ask for a handful of different renderings one after the other. >> As far as I know there are no existing vocabularies which deal >> with user requirements other than basic hardware profiles, but the >> framework for them should work, and it would be, in my opinion, an >> important implementation demonstration. > > Sure, the HTTP headers for this could be expressed as an RDF > vocabulary. I was thinking more about showing an implementation of the system working - a browser that allowed for requests, a server that was answering them, and the ability to change preferences on the fly... An essential component is, of course, a service that is producing these different versions. cheers Chaals -- Charles McCathieNevile charles@sidar.org Fundación SIDAR http://www.sidar.org
Received on Sunday, 12 January 2003 18:45:28 UTC