Re: Moving longdesc forward: Recap, updates, consensus

On Thu, May 5, 2011 at 6:09 PM, Laura Carlson
<laura.lee.carlson@gmail.com> wrote:
> 10.6.1 USER AGENT RENDERING (informative)
> http://www.d.umn.edu/~lcarlson/research/ld-rendering.html
>
> I have updated the 10.6.1 User Agent rendering draft trying to
> incorporate Chaals suggestion to add info on real world
> implementations and Leif's suggestion to add info on iCab's
> contextual-menu cursor. Is it okay?

I'm not sure about this. It seems highly unusual for a W3C
spec to refer to particular implementations by name. I think
it would be better to use their implementations to inform our
examples of how to implement them.

Chaals, did you really mean the spec text should call out existing
implementations by name?

> If there are other aspects of improving longdesc, can they/should they
> wait until after longdesc is reinstated into HTML and be pursued as
> bugs?

I think it's perhaps worth discussing these use cases in the Change Proposal,
in order to demonstrate that @longdesc could be used to provide solutions
to them, if indeed we think it can. Seems to me your CP already does that,
but we could perhaps evaluate @longdesc against other possible approaches?

> CHANGE PROPOSAL
> http://www.w3.org/html/wg/wiki/ChangeProposals/InstateLongdesc
>
> No one has offered concrete suggestions to improve the change proposal
> in the "Moving longdesc forward" thread. Does anyone have any?

I disagree with the claim: "Recent research finds that obsoleting
longdesc specifically breaks the web for over 150 sites in the wild
that are using it to describe images." Obsoleting @longdesc does not
require that implementations drop special behaviours for the @longdesc
element. Anyone actually using @longdesc today must accept that it has
limited support, independent of what W3C
specs say.

I disagree with the "Hidden Meta-data Fallacy" section because I think
it misses the point of the visible data principle, which is that
visible data is less likely to be erroneous because most authors are
more likely to spot visible errors. Ultimately, it seems to boil down
to saying "longdesc" is discoverable, but that's undisputed so I think
the argument is a waste of space that is not going to change any
minds. I'd drop that section.

Under "Recent Research on Assistive Technology", "Thunder in
combination with WebbIE" and "Window-Eyes" are listed as subitems
of "SuperNova/Hal". I think they should be siblings rather than children.

Under "Recent Research on Browsers" I see:

"Longdesc Firefox Extension by Patrick H. Lauke - adds a "View Image
Longdesc" option to the image context menu that activates the link to
the long description (over 23,000 downloads)
"Longdesc for Linux for Firefox 4 by Patrick H. Lauke - allows one to
view image's longdesc from the context menu"

I think these are the same thing and should not be listed twice.

Also, I think we should openly state that Home Page Reader stopped
being maintained half a decade ago, if we're going to cite it
as an example implementation.

--
Benjamin Hawkes-Lewis

Received on Friday, 6 May 2011 01:04:56 UTC