Re: How is aria-interactive different than tabindex=-1

Dominic, we need this for more than just tables. We want to apply it to
lists. We want to apply it to SVG, etc. In SVG, in addition to tabindex we
are looking at other forms of navigation. Please take a look at:
https://www.w3.org/wiki/SVG_Accessibility


Also, we are trying to eliminate the proliferation of roles in the Core
ARIA specification. The group felt that we could reuse grid and gridcell
without having to create a table role and cell role. This of course would
map to the role of table on IA2. So, from an AT perspective they would just
see Table and no "grid".  IOW the mapping layers can handle the cleanup.

aria-interactive is largely intended for static roles like "list". The
group spent a lot of time on deciding whether to have a "table"  role and
it was felt that by simply stating  < div role="grid"
aria-interactive="false" you would make the grid a non-interactive table.
Also, the table would pass the interactive feature of false to all
structural descendants much the same way that role="none" gets propagated.

Rich


Rich Schwerdtfeger



From: Dominic Mazzoni <dmazzoni@google.com>
To: James Nurthen <james.nurthen@oracle.com>
Cc: "W3C WAI Protocols & Formats" <public-pfwg@w3.org>
Date: 05/20/2015 06:33 PM
Subject: Re: How is aria-interactive different than tabindex=-1



I agree with James, a new "table" role is much simpler to understand. I
might like to see "cell" added as a synonym for "gridcell" but to keep
things simple essentially reuse the rest of the grid attributes like "row",
"rowheader", "columnheader". At most that's two new roles.

I also vote to move forward with aria-interactive as proposed. I basically
like all of the text as-is except that I think it should be allowed on more
widget roles than just list and grid. Developers *already* build all sorts
of widgets that respond to keyboard shortcuts when focused, we need a way
to expose this information to screen readers.

On Wed, May 20, 2015 at 4:11 PM, James Nurthen <james.nurthen@oracle.com>
wrote:


  On 5/20/2015 3:57 PM, Matthew King wrote:
        "White, Jason J" <jjwhite@ets.org> wrote on 05/20/2015 02:35:21 PM:
        > Perhaps the ARIA 2.0 discussion will be a suitable
        > opportunity to rework the approach and to establish foundations
        for
        > long-term extensibility. The integration of ARIA with Web
        components
        > could also ease the burden on application authors.

        For ARIA 1.1, the question is whether to:
        1) Move forward with aria-interactive as currently proposed [1]
  JN: We need aria-interactive for things other than tables so I think we
  need to keep it.
        OR
        2) Add table-specific roles to ARIA 1.1.
  JN: Even if we go this route I really don't see the need to add a whole
  bunch of roles for this. Why couldn't we just add role=table which is the
  non-interactive version of role=grid. Both of them can have all of the
  same child roles but these child roles are simply not interactive when
  they are the child of a table rather than a grid. I really don't want
  multiple row, columnheader, rowheader etc. roles in aria but am certainly
  fine with having both a table and a grid.

  Regards,
  James



        [1]
        http://rawgit.com/w3c/aria/matt-action1505/aria/aria.html#aria-interactive



        Matt King
        IBM Senior Technical Staff Member
        I/T Chief Accessibility Strategist
        IBM BT/CIO - Global Workforce and Web Process Enablement
        Phone: (503) 578-2329, Tie line: 731-7398
        mattking@us.ibm.com



        From:        "White, Jason J" <jjwhite@ets.org>
        To:        Matthew King/Fishkill/IBM@IBMUS,
        Cc:        "Gunderson, Jon R" <jongund@illinois.edu>, Dominic
        Mazzoni <dmazzoni@google.com>, W3C WAI Protocols & Formats
        <public-pfwg@w3.org>, Alexander Surkov <surkov.alexander@gmail.com>
        Date:        05/20/2015 02:43 PM
        Subject:        Re: How is aria-interactive different than
        tabindex=-1



        > On May 20, 2015, at 16:43, Matthew King <mattking@us.ibm.com>
        wrote:
        >
        > "Gunderson, Jon R" <jongund@illinois.edu> wrote on 05/20/2015
        12:55:43 PM:
        >> I am wondering how the proposed aria-interactive is different
        from
        >> tabindex=-1?
        >> Both indicate an element has behavior, and the absence of
        tabindex
        >> attribute means no behavior (e.g. aria-interactive=false)
        >
        > Tabindex does not affect mapping. A gridcell in a grid with no
        tabindex specified is still a grid ... it just missing tabindex.
        > Current proposal is that an element with role grid and
        aria-interactive=false would be mapped as a static table is mapped.

        And that’s the contentious point in this discussion.

        To make the matter even more confusing to typical software
        developers (the concern that Jon rightly raised), we also have
        aria-readonly and aria-disabled. Of these, aria-disabled is closest
        in function to aria-interactive=false, except, again, for the
        accessibility API mapping.

        ARIA is already a complex specification. The direction which ARIA
        1.1 is taking makes it even more dependent on subtle semantic
        distinctions that run the risk of leading to errors in the
        implementation of Web applications by well-intentioned ARIA
        non-experts. ARIA is “invisible metadata” as Charles McCathieNevile
        put it in a related context, and this exacerbates the problem.

        To be clear, I think ARIA is much needed, very successful and that
        it makes a highly valuable practical contribution to the
        accessibility of the Web. However, the Web standardization
        community should think deeply and carefully about the risks of
        creating an increasingly complex specification primarily, if not
        solely, for purposes of accessibility, and should strive to find
        ways of making accessible application development possible without
        requiring authors to become specialists in the subtleties of
        accessibility API semantics. Perhaps the ARIA 2.0 discussion will
        be a suitable opportunity to rework the approach and to establish
        foundations for long-term extensibility. The integration of ARIA
        with Web components could also ease the burden on application
        authors.


        ________________________________

        This e-mail and any files transmitted with it may contain
        privileged or confidential information. It is solely for use by the
        individual for whom it is intended, even if addressed incorrectly.
        If you received this e-mail in error, please notify the sender; do
        not disclose, copy, distribute, or take any action in reliance on
        the contents of this information; and delete it from your system.
        Any other use of this e-mail is prohibited.


        Thank you for your compliance.

        ________________________________




  --
  Regards, James


  Oracle
  James Nurthen | Principal Engineer, Accessibility
  Phone: +1 650 506 6781 | Mobile: +1 415 987 1918 | Video:
  james.nurthen@oracle.com
  Oracle Corporate Architecture
  500 Oracle Parkway | Redwood Cty, CA 94065
  Greenping practices and
  products that help protect the environment