RE: WCAG-ISSUE-23 (DavidMacD): We should consider a new "Failure to provide role=presentation on a layout table"

1. >>Failures I don't think should be dependent upon these types of desperate attempts by AT to >>help their users. 
Assistive technologies are designed to  help their users navigate an inaccessible environment.
If all Web content is 100% standards-compliant (with respect to HTML and technology used) AT makers do not have to build heuristics to make best guesses.
It is simply because that these  desperate attempts often offer the next best experience that users are able to plough through content and  get some tasks accomplished.

2. Steve, my first email did give a few examples  of when layout tables are ignored by SRs.
Surely one can concoct all kinds of table markup  and some SRs will treat them as layout tables and some will not.
But as I stated earlier, it is late in the day to introduce this failure because it will render a lot of content that is certified as compliant to now fail.
Instead, suggesting role=presentation as a positive technique  is fair.

3. And as for the  HTML5 table example for the puzzle:
Without the equivalent of a summary, that table is simply not meaningful  and surely fails SC 1.3.1 if one intends it to be interpreted as a data table. 
Supposing for a moment all layout tables are indeed marked up with role="presentation" and I came across this puzzle table, then I'd imagine the author forgot to mark it up as a layout table ... if it had no explanation surrounding it especially with no headers marked up.
This puzzle or a crossword puzzle do use a tabular layout but it is essentially meant to be interactive. 
A form within a table grid  with cells labelled using aria-labelledby/ aria-label as appropriate will make the puzzle accessible and usable.
"I strongly disagree with "Observe the lack of headers, which are not necessary in such a table".
Such tables simply cannot be categorized as data tables.
Yes, a calendar table can have only column header cells marked up and not  row headers in the first column. That still makes it a data table.

4. Is there a failure  for decorative images / images that AT should ignore  that do not have role="presentation" even if they have alt=""? 


On Mon, 6/2/14, David MacDonald <> wrote:

 Subject: RE: WCAG-ISSUE-23 (DavidMacD): We should consider a new "Failure to     provide role=presentation on a layout table"
 To: "CAE-Vanderhe" <>, "Hoffman, Allen" <>
 Cc: "Alastair Campbell" <>, "GLWAI Guidelines WG org" <>
 Date: Monday, June 2, 2014, 12:50 PM
 I think there at least some exceptions
 to that. Such as missing labels and ids on form fields,
 missing alt text, etc...
 Date: Mon, 2 Jun 2014 11:34:17 -0500
 To: Allen.Hoffman@HQ.DHS.GOV
 Subject: Re: WCAG-ISSUE-23 (DavidMacD): We should consider a
 new "Failure to    provide role=presentation on a
 layout table"
 I don’t think it should ever be a failure to not do
 something a particular way.   To pass muster it would
 have to be something where there was NO other way to
 possible do this in any circumstance.     Usually
 when we find something that says “you must do it this way
 or fail”  it is just a verbatim restatement of the SC
 and we don’t list those 
 Most failures are documenting BAD things that
 people commonly do.  Not the lack of doing it right or
 doing it one particular way.
 On Jun 2, 2014, at 10:58 AM, Hoffman, Allen <Allen.Hoffman@HQ.DHS.GOV>
 I like the idea of doing
 this as a sufficient technique myself, while discussions
 take place about if we feel it really is a failure not to do
 this in this way. On the other hand, if we
 can agree at some point we should know if not doing so does
 indicate a failure of 1.3.1.  If
 folks can’t agree to that it can’t be a failure
 condition either. On a subjective personal
 note as a screen reader user I have found that current
 screen readers don’t allow the same level of navigation in
 .css alternatives
 to tabular formatted content, e.g. you can’t read down
 column, in a .css formatted
 tabular format while you can if it’s an actual
 table.  It
 is very frustrating to navigate such content when you know
 the screen reader could read it nicely if it supported
 mapping columns in .css tabular
 materials.  I
 still would not fault a coder for coding it in .css, but would urge my
 screen reader folks to get cracking figuring out how to
 render such materials so they can be navigated
 effectively.  From: Alastair
 Campbell [] 
 Sent: Monday,
 June 02, 2014 11:00 AM
 Guidelines WG org
 Subject: RE:
 WCAG-ISSUE-23 (DavidMacD): We should
 consider a new "Failure to provide role=presentation on
 a layout table" In the
 particular case of layout tables, a consistent question
 since 2008 has been “Can you use layout tables and be
 accessible?”. It is
 actually quite a difficult question to answer, the normative
 text doesn’t say you cannot use layout tables, but there
 is a non-linked advisory for “Using CSS rather than tables
 for page layout”. There is an overall impression that you
 shouldn’t, but nowhere in writing that says you
 cannot. As noted on
 this thread, the impact of layout tables can be minimal, and
 adding this particular failure would change the status of
 many pages. Overall, I
 agree with Loretta, I think an addition sufficient technique
 for using role=”presentation” at least gives a positive
 thing to do if you have to use layout tables. It would
 provide an explicit way of passing SC1.3.1, which would be
 helpful doing audits.-Alastair

Received on Monday, 2 June 2014 17:52:06 UTC