- From: Sander Tekelenburg <st@isoc.nl>
- Date: Sat, 5 May 2007 03:28:43 +0200
At 12:53 +0300 UTC, on 2007-05-04, Henri Sivonen wrote: [...] > Considering all the above, here's my concrete proposal for a solution: > 1) Make style='' a global attribute. For the purposes of document > conformance, make it conforming on all documents regardless how they > came to being. > 2) Include an informative paragraph about how media-dependent use > of style='' is bad. > 3) I make the conformance checker emit a warning (at most one per > document) that paraphrases the informative paragraph when the > conformance checker sees a style='' attribute. > 4) I make the WYSIWYG signature suppress the warning. > > 5) <font> is either made non-conforming or made a strictly inline > element with the attribute color (to avoid <span style='color: ...'> > cruft). I think I would agree with 1 through 3 (Michel Fortin's objects to 3, but the rationale he gives seems to only apply to 4. So I don't understand the objection to 3. But perhaps I misunderstood.) The "at most one per document" restriction under 3 I'm not entirely sure about though. I understand it's counter productive to bug users about issues more often than strictly necessary. But for authors who rely on conformance checkers to find their mistakes so they can fix them, it would be quite unuser-friendly to have this warning not point out every instance of that particular mistake. So ideal behaviour would probably be a single warning, listing all instances. But is that something that can/should be specced? I don't know. Btw, would this be a first for defining something conforming but requiring conformance checkers to emit a warning? Or are there other instances already? I'm not so sure about 4. Several thoughts: - Is somone actually asking for this, or is it just assumed that some party wants this? (At <http://krijnhoetmer.nl/irc-logs/whatwg/20070503>I see the claim that WYSIWYG editor authors want document conformance *and* font/inline style, but I don't recall having seen that actually being requested.) - Why is a conformance checker config option needed to be in the HTML spec? Wouldn't it be better to let people configure conformance checkers themselves? (Compare W3C's CSS 'validator'.) - What is "WYSIWYG", and why only WYSIWYG? In other words, if we should define an opt-out switch for conformance warnings, wouldn't it be cleaner if it were just that ("conformance level=errors|warnings|none"), instead of tying it to 'WYSIWYG' editors specifically? It could be more general, with an optional list of specific warnings to suppress. - How do 'WYSIWYG' editor and host applications communicate? I mean, in the real world, how does an embedded editor actually generate a "WYSIWYG signature" in the head? - As Michel Fortin said: what if 'WYSIWYG' output is changed through some non-WYSIWYG mechanism, or vice versa? Should the WYSIWYG sig be removed by the non-WYSIWYG editor (be it a person or a tool)? If so, what makes us think that person or tool will actually have access to that WYSIWYG sig? And in the other direction, should the WYSIWYG sig be inserted when changing content that doesn't yet contain it? In fact: should the WYSIWYG sig be produced by any WYSIWYG tool, or only WYSIWYG tools that want to suppress this particular conformance cheker warning? Or to put it differently, 4 just seems too complicated to me ;) As to 5, I don't understand what the use is to allow <font color>. The only argument I've heard is that apparently today's HTML->PDF converters ignore CSS. (Well, the affordable ones, that is :)) -- Sander Tekelenburg The Web Repair Initiative: <http://webrepair.org/>
Received on Friday, 4 May 2007 18:28:43 UTC