Re: Summarizing the contentious history of re-opened PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA 1.1) role="presentation" to avoid avoid author confusion

Bryan,

The reason for the role="" is the following:

WCAG 2 would like developers to be able to use aria-label vs. alt for
labeling content. The reason for this is that ARIA has become a common
accessibility API for the Web and it spans more content than just images.
However, to do this we want authors to NOT use aria-label="". That is
inherently bad as it gets misused. Rather when people provide a label we
want it to be be meaningful. However, elements that are indeed
presentational we want to allow them to convey that semantically and do it
in such a way that is with less pain. Allowing for a role="" to have the
same functionality as role="presentation" and having the added benefit of
removing the thing from the accessibility object tree as it is useless will
allow for improved AT peformance - something that alt="" does not do.
James' suggestion is solid.

Regarding the developer applying role="presentation" to every tag. The
person is an idiot and the exception rather than the norm. We can't hold up
good work for people like that. The developer could have but a role of
button on every element, never run an accessibility test tool and call it a
day. We use role="presentation" in hundreds of IBM products and it is used
properly. Like anything you need to educate the developer.

Rich

Rich Schwerdtfeger



From:	"Bryan Garaventa" <bryan.garaventa@whatsock.com>
To:	"Cynthia Shelly" <cyns@microsoft.com>, "James Craig"
            <jcraig@apple.com>
Cc:	"T.V Raman" <raman@google.com>, <jongund@illinois.edu>,
            <jason@jasonjgw.net>, <wai-xtech@w3.org>, <w3c-wai-pf@w3.org>
Date:	01/28/2014 10:13 PM
Subject:	Re: Summarizing the contentious history of re-opened
            PFWG-ISSUE-348: Consider renaming (now actually 'deprecating'
            in ARIA   1.1) role="presentation"  to avoid avoid author
            confusion



The response I've gotten in the past regarding a "" value for an attribute
is 'why bother when it's empty anyway'. It's an ambiguous concept, and for
something like role=presentation that has the potential to screw things up
if mis-used, I'd recommend using an explicit role value that clearly
indicates it's purpose.

For example, I once saw a developer apply role=presentation to literally
every tag as part of a CMS in the incorrect belief that it meant 'to
present', which made the entire site totally inaccessible when it was
rendered.

I think role="span" would have the same problem, because it may be confused

with the span tag by some developers. Another example that I've seen for
instance, is the tendency to put roles that sound like particular roles on
related tags, such as putting role=radio on Inputs with type=radio,
role=textbox on Inputs with type=text and Textareas, etc. I can see the
same
developers putting role="span" on every span tag on the page.

Plus, regarding "" values in general, developers may sometimes put such
attributes within CMSs as placeholder attributes, and this may have a huge
negative impact if it causes screen readers to behave in a particular
manner
simply by doing this.

It would be great if the role didn't specifically match any particular tag,

and also indicated it's intended purpose, EG "void", "null" ? I can
appreciate the difficulty in doing this.

Btw, so many things were discussed during the ftf meeting that I'm having
trouble recalling all of the topics. Do you know if there is a list of
resolution/action items compiled for the total that I could take a look at
to refresh my memory?

Also, for James, is the open source reverse role lookup project public yet
to take a look at? Or is this still in concept phase at the moment?

Thanks,
Bryan



----- Original Message -----
From: "Cynthia Shelly" <cyns@microsoft.com>
To: "James Craig" <jcraig@apple.com>
Cc: "Bryan Garaventa" <bryan.garaventa@whatsock.com>; "T.V Raman"
<raman@google.com>; <jongund@illinois.edu>; <jason@jasonjgw.net>;
<wai-xtech@w3.org>; <w3c-wai-pf@w3.org>
Sent: Tuesday, January 28, 2014 5:32 PM
Subject: RE: Summarizing the contentious history of re-opened
PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA 1.1)
role="presentation" to avoid avoid author confusion


Hmmm....
People don't like alt="" and may not like this for similar reasons.

It can sometimes it can be difficult to teach, especially to people who
don't really know the difference between null, empty string, none, blank
and
space.  There are many such people creating web content, though they are
less likely to deal with aria than with alt.

Other than that issue, I've always been ok with alt="".  Does anyone recall

why people dislike it?


-----Original Message-----
From: James Craig [mailto:jcraig@apple.com]
Sent: Tuesday, January 28, 2014 5:26 PM
To: Cynthia Shelly
Cc: Bryan Garaventa; T.V Raman; jongund@illinois.edu; jason@jasonjgw.net;
wai-xtech@w3.org; w3c-wai-pf@w3.org WAI-PFWG
Subject: Re: Summarizing the contentious history of re-opened
PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA 1.1)
role="presentation" to avoid avoid author confusion

What about role=""? An explicitly empty string for the role value could be
a
synonym for role="presentation"

On Jan 28, 2014, at 4:54 PM, Cynthia Shelly <cyns@microsoft.com> wrote:

> Some of other ideas...
>
> Role=text

FWIW, text is already on the table as a 1.1 role.

ISSUE-435: Consider role="text" to expose elements (and contents) as static

text node
https://www.w3.org/WAI/PF/Group/track/issues/435

> Role=plaintext
> Role=notag
> Role=layout (nice for tables, less sure about other tags) Role=span
>
> I kind of like role=span.  I think it will be really obvious to html devs

> what this does.  It will be a little goofy to devs moving from Windows
and
> other native platform APIs to web, but I think the parallel to HTML will
> be fairly easy to explain to them.
>
> I'd use something else for decorative images.
> Maybe
> Role=decoration
> Role=deco
> Or keep presentation for this use, as it's pretty similar and widely
> deployed.
>
> That could be combined with alt/longdesc/aria-describedby etc. to be read

> on user request, or with aria-hidden to make it silent.
>
> -----Original Message-----
> From: Bryan Garaventa [mailto:bryan.garaventa@whatsock.com]
> Sent: Tuesday, January 28, 2014 1:22 PM
> To: T.V Raman; jongund@illinois.edu
> Cc: jason@jasonjgw.net; wai-xtech@w3.org; w3c-wai-pf@w3.org
> Subject: Re: Summarizing the contentious history of re-opened
> PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA
> 1.1) role="presentation" to avoid avoid author confusion
>
> I'm having trouble understanding how role="inline" would convey to a
> developer that the role would remove the tag from the accessibility tree
> without hiding or removing any child content. Especially since the role
> would be applicable to all elements.
>
> The word 'inline' to me, or 'block', seems to imply that it turns block
> level elements into inline elements or the reverse, which would be an
> incorrect assumption for developers.
>
> Am I missing something?
>
> ----- Original Message -----
> From: "T.V Raman" <raman@google.com>
> To: <jongund@illinois.edu>
> Cc: <jason@jasonjgw.net>; <wai-xtech@w3.org>; <w3c-wai-pf@w3.org>
> Sent: Tuesday, January 28, 2014 8:54 AM
> Subject: RE: Summarizing the contentious history of re-opened
> PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA
> 1.1) role="presentation" to avoid avoid author confusion
>
>
> Jon,
> Borrowing block/inline from CSS  as role values is a good idea; an even
> better idea is to just mirror over CSS state into the accessibility side,

> i.e. make display:inline  create an implicit role="inline"  on the ARIA
> side, rather than asking  authors to write both.
>
> Gunderson, Jon R writes:
>> Another idea is to borrow from the CSS concepts of "block" and "inline".
>>
>> Role="block" and role="inline"
>>
>> This would provide some semantics as to where the "text" content is part

>> of something that stands on its own (e.g. block), versus part of
>> something more (e.g. inline).
>>
>> I know Cynthia Shelley and Rich have talked about concatenating text
>> runs, and this would provide some way to give ATs a hint on how to do
>> that and developers already have some idea what block and inline mean
>> from CSS.
> I am not sure how they would interpret "none", just like the confusion
> over "presentation".
>>
>> Jon
>>
>>
>>
>> -----Original Message-----
>> From: Jason White [mailto:jason@jasonjgw.net]  > Sent: Monday,
>> January 27, 2014 6:12 PM  > To: wai-xtech@w3.org; w3c-wai-pf@w3.org
>> WAI-PFWG  > Subject: Re: Summarizing the contentious history of
>> re-opened
> PFWG-ISSUE-348: Consider renaming (now actually 'deprecating' in ARIA
1.1)
> role="presentation" to avoid avoid author confusion  >  > James Craig
> <jcraig@apple.com> wrote:
>>> Thanks for the feedback Suzanne. Whether or not "none" is the best  > >

>>> replacement is irrelevant. The confusion is not around images. It it  >

>>>  > around the use of role="presentation" on other elements. For
example:
>>>
>>> The following marking: <h4 role="presentation">Foo</h4>  > >  > > is
>>> effectively the same as: <div>Foo</div>  > >  >  > Perhaps
>>> role="generic" would be more descriptive for the uninitiated.
>>
>>
>>
>
>