[sclug] CSS

[Jonathan Leighton]

(Sorry, this was meant to go to the list too)

On Thu, 2005-10-20 at 11:42 +0100, David Given wrote:
> On Thursday 20 October 2005 10:59, Jonathan Leighton wrote:
> > On Wed, 2005-10-19 at 19:05 +0100, David Given wrote:
> > > Unfortunately, CSS is the only game in town if you want web pages. (I
> > > periodically wonder about how feasible it would be to do your layout
> > > using Javascript to modify the HTML code, which ought to let you do
> > > things that CSS can't do, but nothing ever comes of it.)
> >
> > That would -- of course -- be completely and utterly inaccessible. And
> > it's not what Javascript is for anyway.
> 
> If Javascript's not turned on, of course, you get the unmodified original 
> HTML, which should be perfectly readable *without* any Javascript fiddling. 
> (Just like HTML is supposed to be readable and accessible without any CSS 
> fiddling.)

Ok, sorry, I didn't realise you meant that. However, the technique would
have major issues:

 * Incremental rendering would be impossible
 * Added filesize of this chunk of Javascript
 * Javascript is a very slow language
 * Cross-browser Javascript is damn hard

Still, I would find it interesting to see this done. (That's a different
way of saying "I'll believe it when I see it" I guess)

> http://www.cowlark.com/master-index.html. [...]

If it were me I'd probably use Image Replacement to put an image
containing the text in its place. That way I could size the left nav
column in pixels. I'd probably then use a float-based two column layout,
which isn't too hard.

> [...]
> > "most CSS-based column layouts end up having to make some of the columns
> > a fixed size simply to make them work properly" -- no, most fixed width
> > CSS layouts are that way because one or more images are used that would
> > look stupid if the rest of the page was wider than the image. Fluid CSS
> > layouts are perfectly possible.
> 
> Slight clarification here: 'fixed size' includes percentage sizes, em-based 
> sizes, etc. What I meant is that you can't simply let the browser determine 
> the most appropriate size, the way you can if you use tables. Additionally, 
> if you use tables, you can use hacks such as padding images to tell the 
> browser not to make the column narrower than a certain width, which is an 
> extremely useful feature needed to make truly resizable layouts.

A hell of a lot of thought goes into setting out the specs, and there is
probably a reason for everything you have complained about. If there's
not, there's probably an adaquate solution. However, if you feel
particularly strongly that CSS is missing something then I suggest you
put your comments forward to the W3C CSS WG and see what they say. This
argument has probably gone far enough.

-- 
Jonathan Leighton
http://turnipspatch.com/ | http://digital-proof.org/