WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: CSS Units of Measurement

for

From: Jukka K. Korpela
Date: Feb 19, 2004 3:17AM

On Thu, 19 Feb 2004, Tim Beadle wrote:

> On Thu, Feb 19, 2004 at 11:05:59AM +0200, Jukka K. Korpela wrote:
> > Exercise in futility, since anyone who still surfs using NN4 is either
> > interested in content only, not visual appearance, or doesn't know what he
> > is doing and inevitably experiences lots of problems. In the first case he
> > has probably disabled JavaScript, thereby disabling the quasi-CSS support
> > too. If not, should we really punish him by enforcing our guesses on the
> > suitable font size upon him?
>
> Until late last year, our entire company used NN4 (thankfully now replaced with
> NN7). They were interested in more than just content. I think you have made a
> dangerous and sweeping generalisation there.

No, it just falls into the second category I mentioned (though,
admittedly, some of the _users_ might know what is being done to them,
and suffer). No matter what your company does to its own pages, on NN 4
the vast majority of modern Web pages simply don't work at all as
designed, regarding visual appearance.

> > Besides, you never know when your "NN 4 only" stylesheet suddenly
> > affects rendering on other browsers.
>
> Fair point. You would need to make sure that you override every declaration in
> the advanced style sheet, and test, test, test.

Including testing on all browsers that you have no access to, perhaps
because they don't exist yet.

> I don't mean "smaller", "larger" etc; I mean xx-small through xx-large.

So did I.

> Apart
> from IE's incorrect rendering (fixable with Tantek's hack) they have a
> well-defined meaning in modern browsers,

Why would you hack, instead of solving problems, or not creating them in
the first place.

The keyword sizes do _not_ have well-defined meanings. They are defined
as an ordinal scale only, with varying recommendations (and no actual
_definition_) of how they could be mapped to physical sizes. And the only
thing that is reasonably well defined, namely that medium should map to
the user-chosen size, is seriously violated by IE (except on IE 6 if you
use the magic spell to please the Great Doctype Sniffer).

> Isn't that presentational markup, rather than structural?

Yes, it is honest markup that tells that I wish to affect the font size,
relatively. It is not pseudo-structural like <span class="small">.

> What does "big" or "small" mean to a screen reader?

The same as font-size (or maybe potentially a little more, but that's
irrelevant).

> > If you wish to make specific suggestions on relative font sizing, use
> > font-size with percentage value, and avoid getting below 75% or above
> > 150%. Watch out for the effects of nested elements with such settings.
>
> I thought only NN4 did this nesting effect, which I've taken out of the
> equation by using px in an NN4-only style sheet!

No, of course not. If you set font-size in percentage, it by definition
means percentage of the parent element's font size. And browsers mostly
get this right. But authors may not always realize what they are asking
for.

--
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/


----
To subscribe, unsubscribe, suspend, or view list archives,
visit http://www.webaim.org/discussion/