- From: Dave J Woolley <DJW@bts.co.uk>
- Date: Tue, 19 Sep 2000 11:41:35 +0100
- To: "'www-html@w3.org'" <www-html@w3.org>
> From: griffin granberg [SMTP:kuito@kuitoweb.com] > > I have used the validator on the page and > came across border="0" will not work with > frame, frameset.. it does work with IE[damn > microsoft].. but for w3c standards i have > [DJW:] Interestingly I just noticed this whilst trying to convert someone's site to (valid) HTML. It appears that no browser up to IE 5.01 and NS 4.x, at least, correctly implements HTML 4 Transitional Frames. I haven't tried Mozilla yet, and NS 6 won't install because of a an installer fault (wrongly calculates disk space needed). Unfortunately, the rendering model for frames isn't clearly explained and I get the impression that marginwidth and height should really be called padding... for consistency with the CSS model. I suspect that the border controlled by border= is actually margin in CSS terms. The HTML spec does need a rendering model as frameborder and margin* are to do with rendering. I think there are rumours of Frames being put back into XHTML as a module, in which case it would be a good idea to clarify the model. I suspect, as frames have never been part of the reccommended part of the standard, no-one bothered to define their semantics tightly. I assume the intention was that frameborder=0 would suppress the effect you are seeing. (Incidentally, the suggestion of using frameborder=NO introduces an unnecessary standards violation. Also, frameborder must be on frame, not frameset - I've only tried border on frameset.) In terms of validation, I cannot find a DTD that correctly represents IE's behaviour in this respect. The Wilbur derived IE30 DTD dosen't include border=, even though I suspect that IE30 actually supports it. (I'm trying to fix someone else's HTML without making significnat changes to its real behaviour.)
Received on Tuesday, 19 September 2000 09:35:57 UTC