W3C home > Mailing lists > Public > public-html@w3.org > June 2007

Re: Table accessibility (was Re: headers attribute)

From: Henri Sivonen <hsivonen@iki.fi>
Date: Sun, 3 Jun 2007 15:53:57 +0300
Message-Id: <9E0D3B61-898B-467A-A4FA-B18D29673829@iki.fi>
To: HTML WG <public-html@w3.org>

On Jun 3, 2007, at 11:00, aurélien levy wrote:

> After a good sleep i think the test case need more case to be fully  
> revelent so here is what is propose :

What bothers me about this thread is that so far the cases to test  
lack the most obvious case: A table where the table headers are  
marked up as <th> with no scope='' or headers='' anywhere.

The way sighted users work with the obvious case is that for a given  
cell <td> they look up the column until they reach a <th> and to the  
left (assuming ltr text) until they find a <th> and use those as the  
column and row headings. To me, it seems like a no-brainer that for  
practical usefulness, a screen reader should walk tables like this in  
the absence of explicit association. Implementing table walking like  
this is a much better way to improve table accessibility than  
requiring authors to be explicit about the associations in the most  
common and obvious case.

Do any screen readers do this? If not, why not? If not, shouldn't the  
first priority for accessibility experts who want to make HTML tables  
more accessible be lobbying screen reader vendors to bother to walk  
the table model like this?

(I tried to test this in VoiceOver, but I couldn't find documentation  
about keyboard commands querying a table cell for its headers.)

Henri Sivonen
Received on Sunday, 3 June 2007 12:54:15 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:21 UTC