xml2rfc icon indicating copy to clipboard operation
xml2rfc copied to clipboard

Contents of the artwork 'alt' attribute do not appear in output

Open ajeanmahoney opened this issue 2 years ago • 6 comments

Describe the issue

The 'alt' attribute of the artwork element (<artwork alt="descriptive text to help readers understand the artwork">) is supposed to let authors provide a description of a figure so that assistive technology apps can in turn provide the description to readers who are visually impaired in a way that is easier to process (e.g., in audio format). However, when text is included in the artwork 'alt' attribute, it does not appear in the HTML or PDF output.

Code of Conduct

ajeanmahoney avatar Sep 27 '22 21:09 ajeanmahoney

Given that alt text forms a complete alternative to the image, why is this an attribute and not part of <artset>? I realize that you might want different alternative text for a different image format, but it seems to me like the simplest approach is to provide a singular descriptive block, even though that might be limiting.

martinthomson avatar Oct 09 '22 22:10 martinthomson

The suggestion from @martinthomson does seem sensible.

JayDaley avatar Oct 10 '22 10:10 JayDaley

The reason is: history. The alt attribute predates the intruction of artset by at least 15 years.

reschke avatar Nov 06 '23 10:11 reschke

On 2023-11-06, at 11:13, Julian Reschke @.***> wrote:

The reason is: history. The alt attribute predates the intruction of artset by at least 15 years.

… and repeats the most prominent pervasive mistake in RFCXML:

Text doesn’t belong into attributes (because you can’t put markup there).

So like title= became <name, deprecating alt= and using <artset (with a full block mode RFCXML content) seems the way to go.

Grüße, Carsten

cabo avatar Nov 06 '23 10:11 cabo

I like the idea of having descriptive text as an arm of <artset>.

martinthomson avatar Nov 06 '23 14:11 martinthomson

FTR: it just mirrors what HTML has (img/@alt). Over a decade ago, there was an epic WHATWG-W3C war about whether longdesc should be retained; it's gone in WHATWG's HTML5, but there's now a separate W3C spec for it (if I read things correctly).

reschke avatar Nov 06 '23 16:11 reschke