Re: CfC: to maintain "Image Description Extension (longdesc)" Working Draft on the Recommendation track

I object to publishing this "Image Description Extension (longdesc)"
Working Draft on the Recommendation track.

Plan 2014 suggested people focus on providing a better solution to the
use cases longdesc supposedly addresses. But as the chairs have
already pointed out - twice - these use cases can already be solved
with existing alternative techniques:

http://lists.w3.org/Archives/Public/public-html/2010Aug/att-0112/issue-30-decision.html
http://lists.w3.org/Archives/Public/public-html/2012Feb/0058.html

And this has again been confirmed by Charles in the Last Call feedback
noting "It is clearly possible to meet any given use case's
requirements, and even a subset of all of them, with many kinds of
solutions":

http://lists.w3.org/Archives/Public/public-html-a11y/2013Feb/0093.html

PF and HTML A11Y TF have recently again stated their intention to make
longdesc obsolete, but understandably have a "strong concern of
controlling when a longdesc attribute is obsoleted":

http://www.w3.org/2013/10/31-html-a11y-minutes.html#item04

In the meantime, longdesc is still being suggested as a viable
technique, despite the well-known problems with longdesc usage
including basic usability and POUR issues, the current poor level of
UA/AT support, and the hopelessly polluted state of current longdesc
usage:

https://www.w3.org/Bugs/Public/show_bug.cgi?id=22777

It seems there is a much stronger consensus around making longdesc
"obsolete but conforming":

http://lists.w3.org/Archives/Public/public-html-a11y/2012Dec/0030.html

"Obsolete but conforming" would also mean authors using a conformance
checker can be informed via a warning that longdesc is not currently
well supported, and the warning could contain links to alternative
approaches that actually work right now:

http://lists.w3.org/Archives/Public/public-html-a11y/2012Dec/0031.html

I'd like to propose instead:

1. The HTMLWG publish this document as a Note to indicate work - and
discussion - on this feature is finished

2. Reinstate longdesc into HTML5.0 as a "obsolete but conforming" feature

3. Make longdesc "obsolete" in HTML5.1

This approach has the advantage of ensuring authors, implementors and
this Working Group do not spend any more time on a feature that is
known to be problematic for users at the current time, and that is
going to be obsoleted in the near future anyway.

-Matt

Received on Monday, 27 January 2014 18:38:00 UTC