- From: David Woolley <forums@david-woolley.me.uk>
- Date: Mon, 30 Apr 2007 22:01:04 +0100
- To: www-html@w3.org
Edward O'Connor wrote: > I think it's important to realize that HTML5 is a bundle of several > kinds of conformance profiles -- browsers, to be conformant, must > support interoperable error handling in the face of strange input. But > this is orthogonal to the conformance requirements placed on authors. Tbe big difference is that browser developers will, at least at first, honour their part of the standard, but authors will ignore their part. Consequently only the requirements on the browser form the real effective standard. > > For instance, suppose that for reasons of interoperability & not > breaking the web the <tt> element must be supported by browsers (default > rendering with a monospaced font, etc.). It would be entirely reasonable > to simultaneously require authors to not use this element, for whatever > reason -- it's purely presentational, etc. Actually it does have some semantics, although they are rather domain specific and not often encountered. CENTER would have been a better example, and FONT (which *is* retained) an even better one. > > I think much of your disagreement with Lachlan, Tina, is that sometimes > these separate and separable conformance classes are being confused in > the heat of the moment. I think she is just agreeing with Jukka that a standard that requires well defined exception handling in the browser is not, in fact, defining error handling, but rather syntactic sugar in the language, and that authors will code against the browser behaviour, not against a standard that they will never read.
Received on Monday, 30 April 2007 21:01:29 UTC