RE: Cascading Style Sheets

At 8:25 AM -0400 7/8/97, Liam Quinn wrote:

>At 11:50 PM 07/07/97 -0700, Simon Daniels wrote:
>>As has been noted, It's impossible to use CSS to create interesting
>>pages that work well in all the current browsers that claim support for
>>CSS, let alone those that don't. This is because things like CSS layers
>>aren't in IE3, and margins behave differently in different
>Margins and layers should not be critical to the content of a page, so
>such CSS-styled pages should be accessible in all browsers.

I can almost agree, but layers do complicate the question by losing the
necessity - and sometimes the possibility - of linear coherence without
CSS. You can't always go from 3 axes (xyz) to 2 without critical
information loss. But I agree that authors should avoid CSS-dependence
where possible.

Si attaches special significance to the existence of browsers that "claim
support for CSS." Whether they claim support is immaterial, however; what
matters is how well. If they don't support the standard extensively,
according to the best common understanding of its meaning, then its support
is inferior, and that can be dangerous.

I'm drafting chapters for a book in which I recommend using scripts to
serve one or more browser/version/platform-specific CSS files only to
implementations the designer has tested personally. All others get plain
HTML. This will spare designers a lot of grief trying to reconcile the CSS
implementations, especially for more subtle effects, and encourage authors
to start with broadly intelligible HTML, just as print designers begin with
finished copy and graphical assets. It should also encourage designers to
concentrate on the best CSS implementations available on their platform,
and not worry about the often alarming CSS side-effects elsewhere.

>>If anyone can point us to interesting CSS pages that don't use margins
>>or tables and work in all the browsers that claim CSS support then we'd
>>love to see them. To the best of my knowledge they don't exist and never
>What is "interesting"?  It sounds like you want pages that do things never
>seen before, but that's not the purpose of CSS.  A good style sheet should
>go unnoticed by the user.

Well, there are those of us who can't obey traffic signs before analyzing
their composition, so I can't go this far.

To assert that interesting pages will never exist across multiple
implementations is to assert that (a) content is uninteresting, (b) vendors
will never patch the gaps and bugs in their CSS implementations, and (c)
authors will always allow their content to be held hostage to a particular
browser's CSS implementation quirks. If this is what you mean, Si, I hope
your opinions carry little weight with your employer.