WebAIM - Web Accessibility In Mind

E-mail List Archives

More on PDFs with complex headers

for

From: Alan Zaitchik
Date: Nov 4, 2019 10:16AM

I've tried to follow up on various helpful responses from list members.
Thank you all for your suggestions.
What I find, however, when trying out the results with screen readers is
quite baffling to me.

To start with the simplest case, I tried to "flatten" the multi-leveled
column headers by setting the top level to artifact (I artifacted away the
top <tr>, and it no longer is in the Tag tree at all), adding ALT text to
the <th> calls in the row below so that (hopefully) users would get the
same information -- but whereas Voice Over correctly ignored the artifacted
top row NVDA read it anyway. And whereas NVDA recited the ALT text for the
<th> cells that I kept in the tag tree, VO did not. Neither NVDA nor VO
recited the header cells when I navigated to the data cells.


You may ask: why was I doing the above to begin with? The reason is that I
had equally bad results setting up index values in header arrays for the
data cells when I left the top levels of row headers (merged cells) in the
tag tree. I figured I'd be clever and artifact them away, moving their
label text into ALT text for the lower level of headers.

Thanks for any direction you can offer.
A