- From: Charles McCathieNevile <charles@w3.org>
- Date: Tue, 16 Mar 1999 16:11:24 -0500 (EST)
- To: Ian Jacobs <ij@w3.org>
- cc: Chetz Colwell <c.g.colwell@herts.ac.uk>, Wendy A Chisholm <chisholm@trace.wisc.edu>, w3c-wai-gl@w3.org
Wendy's comments marked WC: Chetz' repsonses marked CC: Ian's responses to those marke IJ: And my responses to those marked CMN: Chetz Colwell wrote: > I seem to > remember that there used to be a Technique of using <P> or <BR> in table > cells to to make tables easier (if not accessible) for screen readers. > Could this be recommended until user agents support the association of > headers and cells? Otherwise, an author "must" use attributes that are not > supported by either the main browsers or current screenreaders (I don't > know about other special devices). IJ: I will raise this as an issue with the WG. CMN: As I recall it was to solve a legacy problem when there were a number of browsers which ignored all tables markup, and would therfore run the contents together in a single line. I would be surprised if such a browser was being used by someone who had problems with it, although I suppose those people are out there. > WC:: > >Did any of the authors say why they removed an element or attribute once > >they found it was not supported? Was it because they were frustrated that > >they had "wasted time" implementing it and no one was going to use it? Or > >were they afraid it might conflict with an older browser? > > CC:: > Particpants seemed to be concerned that they would have something 'unknown' > in their page, which didn't work 'now', and they didn't know how a user > might interact with, or whether it might break an older browser. > > WC:: > >If the answer is the second option, then we need to highlight that using > >new features now won't hurt anything (except where noted, e.g. OBJECT). IJ: I'll add a statement to this effect: In general, user agents ignore HTML attributes they don't support and they render the content of unsuppored elements. (I'm not sure why OBJECT "hurts". It was designed not to hurt, i.e., for its content to be rendered when designated objects cannot be operated.) OBJECT hurts in some versions of MSIE 4, which had a really buggy implementation. Although alternative content is still important - the early Netscape implementations didn't render some very common content types (images, text, html) although they would happily render more complex objects. I assume the problem has been fixed now. Part of the problem was that it required a WIDTH and HEIGHT, which can be done as a legacy solution. Also it sometimes crashed. Not much that can be done about that. Charles McCN
Received on Tuesday, 16 March 1999 16:11:27 UTC