On 29/05/2009 12:40 AM, Rob Gallo wrote:
> I actually don't like the idea of described by relations for table headers.
> It's non-standard, and it blurs the line between the table interface and
> relations. In my mind, they have separate purposes.
I'd argue that it does have some precedent in standards; consider, for 
example, the way you can link a cell with its headers using the headers 
attribute in HTML. Relations are designed to specify that two objects 
are related in some way. I'd argue that table cells and their header 
cells are a good candidate for this; they are two objects that are 
"related".

I find the table rowHeader/columnHeader interface to be somewhat clunky 
at best. It would be far nicer if it just returned an ordered array of 
row or header cells for the requested cell, rather than having to return 
a virtual table.

Having said this, it just occurred to me that using described by 
relations is problematic in that you can't determine what is a column 
header and what is a row header without querying the row and column 
coordinates for the header cell, which is pretty ugly. Having to do this 
would make this method equivalently clunky. This could be worked around 
by creating separate relations for column header and row header. 
However, given the shortcomings of IAccessibleRelation, this is perhaps 
not too worthwhile.

-- 
James Teh
Email/MSN Messenger/Jabber: [email protected]
Web site: http://www.jantrid.net/
_______________________________________________
Accessibility-ia2 mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/accessibility-ia2

Reply via email to