Skip to main content

Reminder that support for the `headers` attribute in tables is pants:
• https://issues.chromium.org/issues/40691227
• https://issues.chromium.org/issues/40818082
• https://github.com/FreedomScientific/standards-support/issues/601
• https://bugs.webkit.org/show_bug.cgi?id=235740

Even if `headers` worked, table navigation for spanned cells is still wonky.

So, you know, avoid spanning table cells and headers:
https://adrianroselli.com/2023/02/avoid-spanning-table-headers.html

#accessibility #a11y


Avoid Spanning Table Headers

Spanned table headers are not well supported across screen readers. While you can visually style these all sorts of ways to make the spanning clear, I am focusing on the programmatic outcomes.
Adrian Roselli
#a11y #Accessibility
in reply to Adrian Roselli

Marco Zehe
Marco Zehe
• •
Interesting, I just tried this in latest Firefox on the Mac with VoiceOver as well, and in the first example, the column and row headers are reversed when they are spoken, so instead of saying „ISBN & 13“ like Safari and VoiceOver do, VoiceOver will say „13 & ISBN“. Same with row headers. It will speak the author first and then the region. I’ll defer to @jcsteh to answer whether there is already a bug for this. Since for the Cache The World project, the whole table support has been completely rewritten, and this was after the Firefox version you tested, it is worth revisiting the Firefox results on Windows I think. I no longer use Windows currently, so am not set up to test that myself. But yeah, tables have always been hard to get right.
@Jamie Teh
in reply to Marco Zehe

Adrian Roselli
Adrian Roselli
• •

@Marco
Yeah, good point on the Firefox updates. I have *lots* of Firefox test results that warrant a rewrite.

Time is, of course, the challenge.

@jcsteh

@Marco Zehe @Jamie Teh
in reply to Marco Zehe

Jamie Teh
Jamie Teh
• •
@Marco Firefox walks column and row headers backwards. It should be easy enough to reverse the array before returning, but I'm not sure whether that's always the desired result. In some cases, I would've thought nearest context first would be more useful for user efficiency. @aardrian
@Adrian Roselli @Marco Zehe
⇧