- From: E. Stephen Mack <estephen@emf.net>
- Date: Sun, 27 Jul 1997 16:57:34 -0700
- To: www-style@w3.org
Liam Quinn wrote: >I'm not sure that all elements used should have all (applicable) CSS >properties set. An exhaustive style sheet would override all (non- >!important) user styles, which seems to go against the idea of Cascading >Style Sheets. I recognize that the cascade does not always give optimal >results, but I think that we have to trust it, or at least give it a >chance with some real implementations. We've given it a chance against two current implementations. The problem is that two widely-used browsers, IE 3.x and Navigator 4.01, seem to cascade a BODY declaration so that it overrides the default behavior of almost every HTML element used on a page. I urge you to check my screen shots at http://www.emf.net/~estephen/htmlner/stylebugs.html if you don't have access to Navigator 4.01 or IE 3.02. That the same style sheet results in such different interpretations is discouraging to me. Todd's suggestion is one way around the problem shown here. I think the only reason that David wrote that "prudent authors will have to declare styles for all elements they use" is because of the inheritance problem in Navigator 4.01 and IE 3.02. David's statement in turn prompted me to wish for a look at the default style sheet that IE uses -- which prompted Todd to propose that we work together to generate one. >As an author, I don't want to have to decide every single style for the >user. There are some properties that the user is able to decide better >than me. Whether or not you *use* the default style sheet for every document is up to you. Certainly you'll be able to take out declarations that you want the user to decide. But I think you'll agree that having access to such a sheet is tremendously useful -- first, in light of the behavior of IE 3.x and Navigator 4.01, and second, so that you'll be better able to judge what HTML elements have values that may override your inherited declarations in a correct implementation of style sheets. >Then what we're doing is trying to defeat the cascade. If we dislike the >cascade this much, perhaps it'd be better to attack the problem at the >source and change the cascading model of CSS. (FWIW, I'm satisfied with >the cascade.) I don't think David, Todd or I are against the cascading model. (But then again, I'm not trying to speak for them, just indicate my understanding of their positions.) Instead, I think we're opposed to the current implementation by Navigator and IE 3.x where the inherited values outweigh the UA's initial values. By importing a default stylesheet, we can work around the problematic implementation of inheritence -- putting Navigator and IE 3.0 on (almost) the same model as IE 4.0 pp2, with an unfortunate side effect that user style sheets may be overriden (as you point out). (By the way, I predict that most user style sheets will end up having ! important attached to most declarations.) -- E. Stephen Mack <estephen@emf.net> http://www.emf.net/~estephen/
Received on Sunday, 27 July 1997 19:56:25 UTC