WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: css off

for

From: Jukka K. Korpela
Date: Aug 30, 2009 11:20AM

Despain, Dallas wrote:

> Section 1194.22 (d) Of the section508 guidelines say that
> "Documents shall be organized so they are readable without requiring
> an associated style sheet."

That's a good rule, though not extensive enough. For real accessibility, as
opposed to meeting some "rules" or "standards", a document should also have
essentially the same content with or without a style sheet.

> I thought that the spirit of this law is to make sure that the
> reading order is correct regardless of stylesheet?

I don't think so. Your formulation postulates that there _is_ "the correct
reading order". This may or may not be true, for a document as a whole or
for parts thereof. For example, if you use CSS to place navigation bar last
in non-graphic rendering, you may well actually improve accessibility by
this change of reading order (since e.g. in aural rendering, users probably
prefer to avoid hearing the navigation menu items at the start of each and
every page). On the other hand, if switching CSS on and off changes the
order of things (that aren't in a random order) in unintended and confusing
ways, then you surely have an accessibility problem.

Moreover, the rule really speaks of other things. There are so many things
you can do in CSS that it is easy to create pages that look OK when CSS is
on and fail miserably when it is not. Sadly enough, CSS is even used for
doing content selection that should be performed server-side. There are
sites and services that don't work at all with CSS disabled.

This - and accessibility as a whole - is much more than just being "JAWS
compatible". (And "JAWS compability" isn't even an acceptable criterion.
Rather, JAWS itself is something that needs to be evaluated from
accessibility perspective.)

There a lots of reasons why people may surf with CSS off. Cf. to my "CSS
Caveats (new edition)", http://www.cs.tut.fi/~jkorpela/css-caveats.html

For example, a person might have vision or cognition problems that make
"normal" viewing of web pages very difficult to him, due to widespread use
of tiny fonts, bright colors, insufficient color contrast, etc. Switching
CSS off can be very useful then. Many poorly designed pages become
inaccessible in new ways, but the odds are that they were inaccessible
already.

In fact, people with special needs may really need a viewing mode where
author style sheets are ignored but a user style sheet designed for the
user's personal needs is enforced. This is even more demanding than simple
"no CSS", but roughly speaking we can say that if a page works without any
CSS, it also works with a well-designed user style sheet.

> Our QA department
> is taking it literally and testing functionality of our pages with
> css off (using the "disable styles" function of the wave toolbar)

I think they are doing the right thing, as regards to testing compliance to
the rule you quoted. I hope they go further the way I described and also
check whether the content shown in non-CSS rendering, in addition to being
readable, is essentially identical with the CSS-enabled rendering. (By
"essential" I mean that you could, for example, use CSS to add explanatory
symbols and texts that may help but are not mandatory for correct
understanding of the content.)

> This is causing problems because
> We are using web2.0 style dialogs in our application

Calling them web2.0 style does not change what they are and what they imply.

> similar to these:
> http://developer.yahoo.com/yui/examples/container/container-ariaplugin_clean.html

I'm painfully aware of such approaches, and they are very wrong from
accessibility point of view.

For example, accessing that page on a normal browser with CSS off (easiest
to do in Firefox), I see things like

"Info
Your changes have been saved.
Close"

which make no sense in the context, so they violate the cited rule - the
content is not only different from CSS-styled content but also mumbo-jumbo.

> JAWS respects and doesn't announce things with display:none applied,
> which I think is pretty much essential to designing web 2.0 pages.

If that's what web 2.0 essentially is - and I doubt that somewhat - then we
really need to stay away from web 2.0.

> When you turn styles off in that page, it's hard to navigate

And the page becomes rather incomperehensible, at least in part.

> Is it reasonable to expect these types of
> dialog to work and make sense with css off?

Sense that question does not make. The question is: Are such dialog
implementations accessible? The correct answer is "No".

> I didn't see an
> equivalent wcag2.0 "css off" requirement

So what? Section 508, WCAG 1.0, and WCAG 2.0 are three different beasts, all
with both serious drawbacks and some very good points. Unless you are forced
to do so by law or boss or contract, don't use any of them as a rulebook,
but... well... rather like a guideline, to be taken seriously but to be
ignored when it does not really serve accessibility.

> Is this Section 1194.22 (d) an outdated guideline?

No. It is too limited (does not pose strong enough requirements), but that's
a different thing. You can't wipe away rules just because they make your
options more limited. That's what rules are supposed to do.

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