w3c/wbs-design
or
by mail to sysreq
.
The results of this questionnaire are available to anybody.
This questionnaire was open from 2016-07-07 to 2016-07-14.
14 answers have been received.
Jump to results for question:
See the proposed definition for null roledescription ignored or the diff version for null roledescription ignored.
Choice | All responders |
---|---|
Results | |
yes | 4 |
no | 1 |
concur | 8 |
abstain | 1 |
Responder | Null aria-roledescription treated as if it were not specified and move ARIA text role to ARIA 2.0 |
---|---|
Joanmarie Diggs | concur |
Birkir Gunnarsson | concur |
Stefan Schnabel | yes |
Léonie Watson | yes |
Richard Schwerdtfeger | concur |
Steve Faulkner | concur |
Fred Esch | concur |
Matthew King | yes |
Jason Kiss | concur |
Cynthia Shelly | yes |
Jason White | concur |
Joseph Scheuhammer | concur |
James Craig | abstain |
James Nurthen | no |
In this option, authors could use the img
role with null aria-roledescription
and an aria-label
to provide text equivalent without an apparent role. See the proposed definition for null roledescription on img or the diff version for for null roledescription on img.
Choice | All responders |
---|---|
Results | |
yes | 3 |
no | 5 |
concur | 6 |
abstain |
Responder | Null aria-roledescription suppresses role announcement on img role only and move ARIA text role to ARIA 2.0 |
---|---|
Joanmarie Diggs | concur |
Birkir Gunnarsson | no |
Stefan Schnabel | yes |
Léonie Watson | no |
Richard Schwerdtfeger | concur |
Steve Faulkner | concur |
Fred Esch | concur |
Matthew King | no |
Jason Kiss | no |
Cynthia Shelly | concur |
Jason White | no |
Joseph Scheuhammer | concur |
James Craig | yes |
James Nurthen | yes |
If you have any additional comments beyond your vote above, provide them here.
Responder | Comments |
---|---|
Joanmarie Diggs | I do wish we could have separated out the text+null issue from the new normative statements about AT behavior. They are, after all, independent issues. I'm fine with adding text explicitly stating that null roledescription is ignored and moving the ARIA text/static role to 2.0. If we were being asked to evaluate "Option 1" rather than "Option 3," my vote would be "Yes." Similarly, I'm fine with allowing a null roledescription only on images. The reason I'm voting "Concur" is because I am unsure about the new, and normative, statements about AT behavior. I have to give these changes some more thought. It's not worthy of a "No," however. |
Birkir Gunnarsson | |
Stefan Schnabel | http://rawgit.com/w3c/aria/action2092option3/aria/aria.html#aria-roledescription should include a section that handles "null aria-roledescription suppresses role announcement on img role", when respective question passes with Yes |
Léonie Watson | Would prefer aria-roledescription to be dropped from 1.1, until there has been time to consider the use cases in more detail and find a solution that isn't likely to break more than it will fix. |
Richard Schwerdtfeger | |
Steve Faulkner | |
Fred Esch | |
Matthew King | Allowing null role description even for images is the type of special case hacking that makes ARIA less robust, more difficult to understand, and risky to use. Such special case processing would be difficult to justify even if it were solving an urgent problem that is extremely important to people with disabilities. In this case, the hack is not solving a well defined problem. The primary objective is a minor reduction to screen reader verbosity that many screen reader users believe is actually detremental. In the rare cases where use of such a hack could be appropriately classified as recasting non-text elements as text, there are other means of achieveing the same effect if it is important to the author. |
Jason Kiss | |
Cynthia Shelly | |
Jason White | I think it would be better to allow a value consisting only of whitespace to suppress presentation of the role in the user interface, as this would be less likely due to author error than a null value might be. I also strongly object to citing concerns about potential for misuse as a good reason for omitting or restricting features in ARIA. I think those concerns (and the people who raise them) are well-intentioned, but ultimately, HTML 5, scripting, SVG and ARIA provide more than ample opportunities for authors to make unwise decisions. The Web community has proceeded in this direction without regard to potential misuses, a situation which restrictions in ARIA would not change, while potentially depriving competent authors of the tools they need to create a better interaction experience for users with disabilities. I am also unpersuaded that special cases (e.g., null aria-roedescription on image roles) are at all desirable. In this case, the second proposal would complicate the format unnecessarily, place unnecessary restrictions on the use of null aria-roledescription by confining its effect to image roles only, and address only a single documented use case (images of text). I think such hacks should be kept out of the language. |
Joseph Scheuhammer | |
James Craig | Clarifying: Move text role to ARIA.next rather than specifically to ARIA 2.0. Otherwise this survey is invalid because it asks two questions for each response. To be accurate, it'd need to ask "Move to 2.0?" as a separate question. PS. What happens if everyone votes Yes to both questions? |
James Nurthen | By not allowing an empty roledescription to override the role we are going to end up with authors who want to do this providing even more "creative" hacks such as the one Stefan proposed in the meeting last week or things like aria-roledescription="⇒" which as far as I know isn't read out by any screen readers normally. It will still appear in the accessibility API and the screen reader will still read it. Trying to prevent authors from doing what they want to seems foolhardy - they will find a way to use the technology in order to create the result they need. However, by preventing them using sane "hacks" like aria-roledescription="" you will force them to use crazy hacks. |
The following persons have not answered the questionnaire:
Send an email to all the non-responders.
Compact view of the results / list of email addresses of the responders
WBS home / Questionnaires / WG questionnaires / Answer this questionnaire
w3c/wbs-design
or
by mail to sysreq
.