E-mail List Archives

Re: "Until user agents..."

for

From: Jukka K. Korpela
Date: May 24, 2004 7:50AM


On Mon, 24 May 2004, leo wrote:

> As many of you are no doubt aware, there are a number of WCAG 1.0
> checkpoints that being with: "Until user agents..."

You raise an important question, especially since many authorities refer
to WCAG 1.0 as a norm, often completely ignoring the "Until user
agents..." condition. Unfortunately the authorities may lack all
understanding of the subject matter; they just pledge allegiance to WCAG
1.0, directly or indirectly via subscribing to another authority's
position. Or they stare at reports from purported accessibility checking
software, which naturally knows nothing about real-world conditions.

> They are 7.1 (flickering), 7.2 (blinking), 7.3 (moving content), 7.4
> (refresh), 7.5 (auto-redirect), 10.1 (spawned windows), 10.2 (explicit
> form associations), 10.5 (adjacent links), 1.5 (redundant text links for
> client side image maps), 10.3 (side by side text), 10.4 (empty form
> controls).

Item 10.4 is especially bad. It was probably a bad idea from the beginning
(the drawbacks weighing more than the possible benefits), and especially
now, but it _is_ there, people read it, programs check it, and authors
reduce the accessibility of their pages when doing their best in what they
regard as conformance to accessility requirements.

> It seems to be the consensus that 1.5, 10.3 and 10.4 no longer apply,

1.5 is surely still an issue; just open a page with a client side image
map using any normal graphic browser with image loading disabled.
The image map is virtually useless. More on this:
http://www.cs.tut.fi/~jkorpela/html/mapalt.html

10.3 is still relevant, though not as much as it used to be. And it may
gain relevance as small devices become more popular. But if layout is
achieved using CSS, then there's usually no big problem: when properly
designed, the design linearizes automatically when the page's style sheet
is ignored.

Regarding 7.1, 7.2, and 7.3, there's not much progress in user agents, I'm
afraid. However, authors might consider using e.g. blinking or rolling
(marquee) presentation, _if_ they provide for a page-specific method for
freezing it in sufficiently many cases. There's one approach described at
http://www.cs.tut.fi/~jkorpela/styles/notice.html
It works via CSS and JavaScript (or "DHTML", if you like), so it surely
does not work universally. But one might argue that adding, say, moving
content that can be wiped away with a simple user action in the great
majority of cases isn't a big problem in accessibility. I would still say
that it violates 7.3 (though its explanation might be read as accepting
the approach I describe).

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