Re: Selectors+accessibility (the XML threat)

On Sat, 23 Aug 2003, Silas S. Brown wrote:

> The vast
> majority of users with special needs have more "minor" needs
> such as low vision, dyslexia, or certain medical conditions
> that make certain colour combinations (or movements)
> painful.  For these users, screen-reading technology is
> usually "overkill" and prohibitively expensive.  A
> well-written user-supplied CSS would be much more effective
> and also cross-platform.

Yes, I think I see what you mean, and you make an important point.
User style sheets could even now be used much more, for accessibility
reasons, or just for convenience. Quite often a simple user style sheet is
used just to _prevent_ something from taking effect, such as to prevent
author's attempts to set font size very small.

> Often it is not necessary to infer the complete logical
> structure of the document; it's merely necessary to say
> "don't let anything blink"

Well, that should be easy, in principle, in current CSS.

> or "disable floats"

That would be technically simple, in a sense, but the overall effect of
floating varies a lot. I'm afraid we can't really predict what it implies
if we disable floats.

> or "use this
> alternative colour instead of italic because I can't read an
> italic font".

(Or because the browser runs in an environment where italic fonts are
unavailable.) OK, now I see the idea. Effectively, I would say, you would
like to set CSS rules that depend on the settings of properties other than
the one being set. For example, just to map one setting to another.

This would be a major change in CSS philosophy from the relatively simple
principles of CSS1 and CSS2. Given how much it seems to take to have even
decent CSS1 support universal (at least in the sense of not spreading
seriously buggy implementations around), I'm afraid anything significantly
more complex, and hence more difficult to implement, would take many years
to become common, even if a specification were issued today.

But I think that if the idea is to be incorporated into future CSS
versions, it should be based on the structure of CSS rules, rather than
introducing new pseudo-elements or pseudo-classes. What would be needed is
a rule that says (globally or for the elements that match a given
selector, or set of selectors) "if property P has value X, then use the
initial value for P instead, and set property Q to value Y". (This would
involve primitive programming, in a sense, and would take CSS a little
closer to a programming language. I don't think this should be an
obstacle, but others might.)

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

Received on Saturday, 23 August 2003 13:21:33 UTC