W3C Proposed Recommendation: longdesc

Jonas Sicking jonas at sicking.cc
Wed Jan 7 11:15:51 PST 2015


On Tue, Jan 6, 2015 at 3:13 PM, L. David Baron <dbaron at dbaron.org> wrote:
> (I'm not happy about this spec; for a good description of why, see
> http://lists.w3.org/Archives/Public/public-html-admin/2014Aug/0028.html .
> I'm also under the impression that they're using Mozilla's
> "implementation" of it as support for the spec, which is rather
> galling considering that a big piece of what led to that
> implementation was an online harrassment campaign against a Mozilla
> UX designer to get the feature accepted.  I'm not sure how much it's
> worth fighting it at this point, although probably a first step in
> that fight would be to remove our implementation.)

I tried winning this fight, but found it not worth my time. The
problem was that even after getting this feature removed from W3C
specs once, the proponents of it simply harassed people again and
again until it got added back to the HTML spec. I guess it got removed
again and yet again they were able to get it back.

The W3C process is clearly not robust enough that we can prevent this
type of stuff form making it into a spec. So I agree that the best
tool we have at our disposal is to simply refuse to implement, or in
this case remove our, as I understand it largely useless,
implementation.

It's sad that people are spending time on features like this when
there are far bigger accessibility problems with the web platform. The
fact that native platforms, especially iOS, has caught up and
surpassed the web when it comes to delivering accessibility "built in
by default" is a sad state of affairs. The fact that the web was based
on a semantic language like HTML was always supposed to deliver a
strong accessibility story.

Sadly it has for authors become easier to deliver beautiful websites
if they simply create an endless pile of <div>s than if they actually
use semantic markup.

That would be a great problem to try to attack.

/ Jonas


More information about the dev-platform mailing list