(cross posted from [email protected], as the discussion is happening in
both email lists)
Sorry this isn't attached to any of the previous threads on the topic.
I'm not part of the HTML WG and can't reply directly to the emails.
John Foliot asked for links to the research that led to the decision to
drop @summary. Henri has posted links to the IRC entries, and they're
good, but they lead to the actual information. Mark Pilgrim, Ian
Hickson, and Philip Taylor all provided information, so if I miss
anything, I hope they'll provide correction:
There seems to be two separate sources of data that led to the decision.
The first was a set of videos discussing all of the accessible markup.
These can be found at http://www.cfit.ie/html5_video/. Note the sound
quality is bad, and for someone with some loss of hearing, like myself,
they were very hard to hear.
The key to the second set of research can be found in a WhatWG blog post
that Mark Pilgrim did a few years back, at
http://blog.whatwg.org/the-longdesc-lottery. We find out in this post
that Ian has been performing tests of data in the Google index, to look
for instances of @summary @longdesc et al use, and then examine how
they're being used. It's from this analysis that Ian, and presumably
other members of the WhatWG effort, formed their decision to remove the
accessibility markup.
I have a "tweet" into both Ian and Mark to ask what is probably the same
question many of you have: is this same raw data accessible to everyone
to perform tests, or evaluate the testing method? Because if not, then
of course the test results have to be invalidated. Tests that can't be
reproduced outside of Google cannot be independently verified and are
therefore, nothing more than additional anecdotal evidence.
Philip uses the dmoz directory for testing, but again, this site is not
representative of the web, as a whole, as the entries in dmoz tend to be
self-selecting, and therefore not an especially good test subject. Not
if we're really looking at "web scale"--a term that was referred to
several times in the IRC discussion.
The other concern expressed in the IRC, rather emphatically, too, if one
looks at the exclamation points, is the fact that we don't see
widespread use of @summary after ten years! Half the web! (Those are
more or less direct quotes from the discussion.)
Of course, I'm not an accessibility expert, just an interested
bystander, but I've noticed that--and this is unfortunate--changes in
general behavior in order to provide support for a minority, in this
case those with physical challenges, tends to happy very, very slowly.
In fact, rather than look at the web for determining expectations as to
the eventual success of @summary et al, one should look at other
accessibility efforts, the time it took for them to take hold, and the
problems they have today. Efforts such as handicap parking, ramp access
to sidewalks and commercial buildings, walk signs that provide an audio
signal and so on. Again, though, I'm not an accessibility expert but
measurements based solely on markup usage ignore the larger issue of
society's reluctance to expend the effort in order to assist those who
need the assistance.
Another individual in the IRC discussion mentioned about using the ARIA
describeby capability. I don't know if a discussion about providing an
alternative approach has been broached. But mention of this also seemed
to touch on concerns that both Mark and Ian had: that @summary was being
pushed when other options would be better. Or at least, less
problematical. Perhaps the accessibility folks can specifically address
this.
All in all, the IRC discussion was very informative. I know that IRC
isn't necessarily 100% accessible, and not altogether easy to use, but I
would recommend that other folks jump into the #whatwg discussions to
ask more direct questions in the future. It's not the proper venue to
make decisions, but it is a good place to address some
misunderstandings, and find out additional info.
Shelley
