- From: Daniel Dardailler <danield@w3.org>
- Date: Tue, 19 Aug 1997 14:12:17 +0200
- To: w3c-wai-wg@w3.org
Chuck Letourneau wrote > If the officially sanctioned W3C Page-Authors' Guideline does not > deliver such information, there may continue to be a need for > supplementary guides like Trace or Starling to fill in the gaps. The issue of scope vs. time is not new (see msg http://lists.w3.org/Archives/Public/w3c-wai-wg/1997AprJun/0078.html in the archives and the minutes of the June 18th conference call at http://www.w3.org/WAI/group/970618call.html) but since this thread has started I've been thinking of new ways to adress it. One new element, at least in my opinion, is the vigor with which visually impaired users themselves, like Raman, Jason, or others at the Boston meeting (Judy Dixon, Steve Tyler) are advocating the "let's focus on the document-encoding" view (or forget the idiosyncrasies of specific user agents, as Jason put it). So on one hand, I'm totally in favor of dropping all these "one link per line" and "BR in TD" element from the WAI Markup guidelines and concentrate on the quality of the source itself, and nothing else. In a sense, that's working for the future, which is a position I'm happy to stand for. On the other side, as Chuck presented it, there are people that want to do more than what's just needed, and want to serve the largest population of today's Web (and therefore for whom knowing the idiosyncrasies and how to cure them at a point in time is important). These people shouldn't be left out the WAI. It's now clear that we want to separate the "pure" guidelines (that apply to the markup out of a particular browser context, like ALT text presence) from the agent-dependent guidelines, the question left is whether we want to present this split using: 1- footnote/subsections in the overall "pure" document 2- as a separate document still edited by the WAI group 3- as a separate document edited by some other groups I think we're heading toward 1 after the August meeting, but I'd like to propose that we do 2 instead (Gregg, how about that for a moving target...). The issue, as Raman put it, is one of perception, more than anything else, and one of our requirement should be that content providers (mainly) and authoring tool makers feel comfortable they can easily implement our guidelines when they look at them, so the shorter and more focused the document, the better. To formalize a little more using Raman model: > WWW access is a function of the following triple: > (document-encoding, user-agent, access-agent)_t I personally think it is (document-encoding, user-agent)_t (I don't see why we need to separate access-agent for the overall user-agent in the theory, and I also like to point out that document-encoding include anything you get from the server). And what we could achieve is to make WWW access *guidelines* a even simpler function: (document-encoding)_t where the only think we care about timewise are the versions of HTML in support.
Received on Tuesday, 19 August 1997 08:12:21 UTC