W3C home > Mailing lists > Public > public-html-a11y@w3.org > August 2012

Re: why are we pursuing this idea? (was: Implementation Details request on Issue 204 Decision)

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Wed, 22 Aug 2012 02:23:01 +0200
To: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Cc: Steve Faulkner <faulkner.steve@gmail.com>, John Foliot <john@foliot.ca>, Maciej Stachowiak <mjs@apple.com>, public-html@w3.org, HTML Accessibility Task Force <public-html-a11y@w3.org>
Message-ID: <20120822022301342526.0daef070@xn--mlform-iua.no>
Benjamin Hawkes-Lewis, Wed, 22 Aug 2012 00:46:18 +0100:
> On Tue, Aug 21, 2012 at 6:21 PM, Leif Halvard Silli:
>> Benjamin Hawkes-Lewis, Tue, 21 Aug 2012 17:11:33 +0100:
>>> On Tue, Aug 21, 2012 at 4:59 PM, Leif Halvard Silli:

>>> Doubtless some will also be surprised if @hidden content _is not_
>>> presented. I have a suspicion this may be the smaller group. If we had
>>> stuck with the original name of @irrelevant, this would have less
>>> surprising.
>> 
>> Well, here I strongly disagree:
> 
> You mean you disagree with my suspicion?

Yes. Your above claim was that only few - the smaller group - would be 
surprised if @hidden causes e.g. an aria-describedby to not have effect 
when its points to hidden content.

>>   a) ARIA allows to point to hidden content and this - give and
>>      take - works well.

I should have said that I, by 'hidden content', did not refer to 
specifically to @hidden.

> I've only seen one attempt at doing this in the wild, and the author's
> expectation was that hidden content would be excluded when referenced.

If he referenced it with e.g. aria-labelledby, then may he needs to 
sharpen his reasoning.

> The examples provided in the Change Proposals are a mixture of the
> broken (e.g. an ASCII art image is left without a defined accessible
> name) and the nonsensical - and that's _after_ various other mistakes
> I pointed out were corrected. So … what makes you think it "works
> well", assuming that you're factoring authors getting the markup right
> into the equation?

Sorry, I did not mean @hidden. I mean the *concept* of referring to 
variously hidden content. For instance, as I said @alt is hidden, and 
@aria-label is hidden, and @aria-labelledby may point to something 
hidden etc. Thus, there is a pattern there. 

However, of course, ARIA can be difficult to get into. It took a while 
before I started to get it.

>>   b) A11Y annotations very often are hidden from most users (e.g.
>>      think of the content of the alt attribute). Or such things
>>      as placing content off screen etc.
> 
> @alt and off-screen content are displayed to users in the
> comparatively familiar cases where you have images or stylesheets off
> or they don't load.
> 
> That's very different from @irrelevant.

I agree. However, the concept that something that is - for the time 
being - hidden/invisible, can still be available to A11Y users, is well 
known.

And also, as you know, aria-label/labbelledby/describedby are not - 
unlike @alt -  affected by whether images load or whether CSS is 
enabled or disabled. So already when using the 
aria-label/describedby/labelledby attributes, one is in danger with 
regard to risk for creating content that has an imperfect fallback in 
e.g. text browsers. So, if referring to content that is @hidden in and 
by itself is such a big problem, then we should perhaps go deeper and 
look at aria-label/describedby/labelledby as well.

In a summary, I think that, with aria-label/describedby/labelledby and 
@hidden, one is already in trouble. And I think that the purpose of 
ARIA very much is to make something that works when the Web site is 
consumed the way it is intended to be consumed - with scripting, CSS 
and everything. To make it *also* work in a text browser thus requires 
that one think of specifically that problem.

>> Thus, 1) It is hypothetical that it could be made not to work.
> 
> What do you mean? I didn't say anything about making things work, I'm
> just talking about author expectations.

OK. Sorry if that was a bad point ...

>>       2) It does not really contradict what authors are used to.
> 
> Authors aren't used to a11y markup at all. (In my day job I interview
> plenty of people employed as frontend developers who don't know the
> <legend> element, who don't understand how <label> is associated with
> form fields, and to whom ARIA is barely more than a rumour.)

Weird interviews you make. :-D Yes. I can imagine.Myself, I think that 
I would recommend using display:none rather than @hidden, as 
display:none does not have the additional aria-hidden=true meaning and 
is thus more flexible when creating JavaScript-free interactivity via 
CSS and hyper links.

But my point was very small: Like I said above, authors often know 
about @alt. And they know that @alt does not render - unless unless ... 
And thus they are used to the concept that invisible content can be 
useful to e.g. unsighted users.
-- 
leif halvard silli
Received on Wednesday, 22 August 2012 00:23:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 22 August 2012 00:23:35 GMT