RE: Action-1293 Proposal

thank you. I will take a look.


Rich Schwerdtfeger



From:	Cynthia Shelly <cyns@microsoft.com>
To:	Dominic Mazzoni <dmazzoni@google.com>, Alexander Surkov
            <surkov.alexander@gmail.com>
Cc:	Matthew King/Fishkill/IBM@IBMUS, Joanmarie Diggs
            <jdiggs@igalia.com>, PF <public-pfwg@w3.org>, Richard
            Schwerdtfeger/Austin/IBM@IBMUS
Date:	03/11/2015 02:28 PM
Subject:	RE: Action-1293 Proposal



I made a proposal to handle this for lists as part of my Web Accessibility
Properties and Actions proposal at TPAC.  It could easily be generalized to
tables as well.

https://rawgit.com/cyns/wapa/master/wapa.html

From: Dominic Mazzoni [mailto:dmazzoni@google.com]
Sent: Wednesday, March 11, 2015 10:20 AM
To: Alexander Surkov
Cc: Matthew King; Joanmarie Diggs; PF; Richard Schwerdtfeger
Subject: Re: Action-1293 Proposal

I still don't understand what this line is supposed to mean:

<tr aria-rowindex="100" aria-rowcount="50"><tr> <!-- 49 rows are not yet
loaded -->

If the intended meaning of this line is that "there are 49 rows following
this one that aren't loaded yet", then I don't see how that's useful
information to annotate. I'm not aware of an existing native accessibility
API that has a way to express this, nor am I aware of any screen reader
that announces anything like this. Rather, most APIs and screen readers
primarily concern themselves with (1) the coordinates of the current cell,
(2) the span of the current cell, and (3) the total number of rows and
columns in the table.

Everything else is just implied or discovered - for example if you navigate
down from row 1 and reach row 100, clearly the rows in-between were missing
- but that's implicit, not explicit.

- Dominic

On Wed, Mar 11, 2015 at 10:01 AM, Alexander Surkov <
surkov.alexander@gmail.com> wrote:
 I don't see a reason preventing rowcount from working on the table element
 (I guess same for colcount attribute) but I let Joanie to comment in case
 if I miss something.
 I would argue that spanned cell and missed rows/columns are semantically
 different and shouldn't be used to substitute each other.

 On Wed, Mar 11, 2015 at 12:38 PM, Dominic Mazzoni <dmazzoni@google.com>
 wrote:
  On Wed, Mar 11, 2015 at 5:38 AM, Alexander Surkov <
  surkov.alexander@gmail.com> wrote:
  yes but not always though. If that was true then there's no reason of
  putting those attributes on cell or row elements. As I understand a table
  may be loaded in number of parts (Joanie's point iirc), for example:
  <table>
    <tr></tr> <!-- 98 rows are missed-->
    <tr> aria-rowindex="100" aria-rowcount="50"><tr> <!-- 49 rows are not
  yet loaded -->
    <tr aria-rowindex="151" aria-rowcount="20"></tr> <!- 19 rows are not
  yet loaded -->
  </table>

  I'm not sure I understand that example. If the user lands on row 151, how
  would you have that described by AT? Aren't there 170 total rows in the
  table?

  What would be wrong with this?

  <table aria-rowcount="170">
    <tr aria-rowindex="100"><tr> <!-- 49 rows are not yet loaded -->
    <tr aria-rowindex="151"></tr> <!- 19 rows are not yet loaded -->
  </table>

  In that example above, the index of each row, plus the total row count in
  the table, is sufficient to totally describe what rows are present and
  what rows are missing.

  Alternatively, if you want a placeholder for missing rows, I think we
  should use rowspan:

  <table aria-rowcount="170">
    <tr aria-rowindex="1" aria-rowspan="99"><tr> <!-- 99 rows are missed
  -->
    <tr aria-rowindex="100"><tr>
    <tr aria-rowindex="101" aria-rowspan="49><tr> <!-- 49 rows are not yet
  loaded -->
    <tr aria-rowindex="151"></tr>
    <tr aria-rowindex="152" aria-rowspan="18><tr> <!-- 18 rows are not yet
  loaded -->
  </table>