RE: Screenreader says....

The reality is that it’s a bug, not a feature.

Screen readers and verbal magnifiers read things oddly because they have been DESIGNED to expect bad code.  It’s the reason LABEL, IDs, SUMMARY, and ALT are all able to be read on an image inside a table.  Only one is correct, but the variants MAY work with some AT.  (smile)

The reason there’s no documentation on this is because the W3C WCAG  (and attendant supporting techniques and syntax standards like ARIA) are the standard, and thus while AT will support people who want to hear more or less info, the expectation is that if you code it right, it will read the way the customer has chosen.   Some people just ignore images, just as some people ONLY look at images when reading a site.

Final note:  this is why you check your code first, THEN have a trained AT user or professional AT tester give it a run.  AT is sometimes too smart for its own good, and verbosity and other settings can cause a “miss” on things you coded correctly.

Regards,
Mark D. Urban
CDC/ATSDR Section 508 Coordinator
Office of the Chief Information Officer (OCIO)
Office of the Chief Operating Officer (OCOO)
Murban@CDC.gov<mailto:Murban@CDC.gov> | 919-541-0562 office
[cid:image001.png@01D39E67.CE8948C0]

From: Alastair Campbell <acampbell@nomensa.com>
Sent: Thursday, December 19, 2019 5:04 AM
To: caroline <woodward.caroline@gmail.com>; w3c-wai-ig@w3.org
Subject: Re: Screenreader says....

> I was wondering if there is some sort of documentation out there that helps writers and developers know what the screenreaders will say when certain code is used. For example when <li> is used JAWS automatically tells the user, 1 of 2...

I would avoid that if possible, screenreaders deal with elements differently and even within a particular version the user can personalise how things are read out.

Instead, I’d make sure the writers understand what the semantics of the available mark-up should be used for.

For example, <img> tags are treated as inline content, so punctuate appropriately. E.g.
<p>I <img src=”heart.png” alt=”love”> New York.</p>

That would generally be read out as: “I love (image) New York”. However, declaring the image might be skipped if the user has set a low verbosity setting, so it still makes sense without that.

If it were a picture of something that is separate from the text, I’d add a full stop at the end of the alt text.

In the case of bullets, I’d give a general statement that: If bullets or numbering are used visually then the appropriate structure must be used. That structure allows screenreader users to skim the content quickly (e.g. skip the list) and get an idea of how many items there are at the start of the list.

BUT, you should not be definitive about how things are read-out because you cannot predict or control that directly, and you don’t want the writers to rely on that single interpretation.

Kind regards,

-Alastair

Received on Thursday, 19 December 2019 13:56:21 UTC