RE: Alt verbosity [was: Example canvas element use - accessibility concerns]

Benjamin Hawkes-Lewis wrote:
> > John Foliot wrote:
> 
> > Specifically I point to your
> verbose
> > @alt values, both in this example as well as examples in the current
> draft
> > specification.  However countless others have noted on numerous
> occasions
> > that when it comes to images, many (most? all?) non-sighted users
> wish to
> > choose between a verbose or terse description, similar to 'glancing
> at an
> > image' vs. 'studying an image'.  This is an important distinction
> which
> > seems to be lost here.

Whether or not the survey is an accurate barometer of the question at hand I
will concede, although it is *some* data, and is more than is being offered
for the counter argument for verbose @alt text. However the main point I am
attempting to make here is that often an overly verbose @alt value is
detrimental to the end user.  Consider:

...alt="The White House"
...alt="Photo of the White House"
...alt="A Black and White photo of The White House in Washington DC taken by
Ian Hickson on February 31st, 2010. It shows the West Wing of the White
House taken at approximately 7:30 PM and you can clearly see the lights
being left on in the various rooms."

Now, in some instances, those particular details might be important to some
users.  Being a sighted user, it will likely be evident to you that some
lights are being left on.  You might take that information in 'en-passant'
and process that cognitive bit as significant, or not.  You have the
'luxury' of deciding whether to study that photo for every nuance and detail
(which rooms have lights left on?  Who's rooms might they be? How many rooms
are lit versus how many are not, etc.) You as a user (sighted or not) may or
may not care about the date and time of the photos creation: should it then
be part of the @alt text? (As an aside, there has been some collateral
discussion around including EXIF data as part of the 'description' data of
photos, and how that might be exposed and processed by user agents - it's an
interesting and useful idea IMHO)

What I am suggesting however, is that on first pass, this amount of detail,
while certainly useful (and oh that I wish was always made available to the
non-sighted) is likely too much.  I am asserting that alt="Photo of the
White House" is likely enough for the majority of non-sighted users most of
the time, and I have asked informally of others reading this thread to
comment - so far none have taken the opportunity to do so (sadly).

> > While not 'conclusive', WebAIM's survey results (and interpretations)
> of
> > screen reader users state:
> > "The tendency toward the briefer alternative [text] also increased
> slightly
> > with screen reader proficiency"
> > [http://webaim.org/projects/screenreadersurvey/#images]
> >
> > Given those findings, I will suggest my 'opinion' is based upon more
> than my
> > personal preference, and seems to directly contradict your opinion.
> 
> John, I doubt the WebAIM survey results are worth citing as evidence on
> this particular topic.
> 


If the WebAIM survey is neither precise enough, or broad enough to drive
home this conclusion, then perhaps we need another survey.  If this is the
case, then either my friends at Utah State might run another survey, or I
can sponsor one here at my work.  What other specific questions do we need
to be asking? (I have some ideas, but am quite willing to open this up to
other suggestions - in fact value them).  Thing is, it appears that hard
data is hard to come by, which again leaves us with the even thornier
problem of having the HTML5's 'opinion' guide the specification, rather than
informed decision making.


> 
> I doubt there's an advantage for short text that leaves out useful
> detail when you're just reading through a page lineally. Why summarize
> an "img" element but not a "p" element?

Really?  How do you arrive at this conclusion?  Are you a daily user of
screen reading technology?  Have you asked daily users how much detail they
want for every image?  Do you not think that having a choice between a brief
description and a detailed description is a better offering than imposing a
long-winded description every time?  If you answer yes, why?

Leaving aside the "thorny question" of the value of WebAIM's survey results
and their validity, what research do you or the HTML5 working group have
that establishes the certainty that a verbose @alt value is a better
solution?


> Most of the time, you're going
> to need to expand both to understand the page, in which case you've
> just given the user more to process.

This is a bold statement.  Most of the time the user will need to expand an
image's description to better understand it?  Please give a concrete
example: I provided 3 examples at the start of this response, 2 short and
one verbose.  I wonder aloud how many non-sighted users would need an
expansion on "Photo of the White House"?

I am not disagreeing that there should not be a mechanism to expand upon a
visual assets initial description - hell I'm still arguing for @longdesc -
but to presume that all non-sighted users should drink from the Fire hose is
(Mr. Hickson) what is frankly very rude - it presumes that you as the author
knows better how and what the end user will be doing with your imagery -
that all users are a monolithic mass.  It's a 'we know best' attitude that
prevails many of the accessibility pronouncements from the WHAT working
group.

There has been so much electronic ink spilled over @alt in the past 18+
months that finding previous references and quotes is a monumental task,
however I do recall that in at least one thread, a now blind user (was it
Leonie from Nomensa?) commented that users who gradually became blind valued
a fuller contextual explanation of the image than those users who have never
seen - that it helped the first group better process information based upon
their past experience.  However, it was also stated that blind-from-birth
users did not value the same depth of detail, as 'visualization' for them is
a very different concept.  I again reference the difference between glancing
at a photo versus studying a photo - they both involve processing visual
information, but under most circumstances sighted users will glance at an
image before deciding to further study an image.  If this is indeed a
truism, then why would a non-sighted user not want to process data in a
similar fashion?  


> 
> However, I can imagine moving from "img" to "img" trying to find the
> image you want, and that short titles for those images could be useful
> so you didn't need to listen to a full alternative.

 A good use case, yes.

> But is "alt" -
> intended for alternatives - really the best attribute to use for such
> short titles? Might the "title" attribute be a better way to provide
> such them, or "aria-label" and "aria-labelledby" from ARIA?

Well, let's return to paved cow-paths shall we?  Most images on the web
today have brief descriptors for @alt, and so both users and authors have
become accustomed to using @alt in this fashion. Sadly, there are probably
too many documents out there where @title and @alt are the same value (if
for no other reason than to ensure "tooltips" in the major browsers), that
making the use of @title alone in the fashion you suggest a missed
opportunity (so let's learn from past mistakes shall we?). 

@longdesc, while maligned and abused (or simply under-appreciated) is being
contemplated for dismissal, it is in fact a good mechanism for the described
action/use-case.  ARIA's "described-by" is for the most part a re-creation
of the feature-set, but if a fresh coat of paint and a new name gets things
on track, hell call it a "thingamabob-arama" and just start using it.  But
to try and re-cast @alt as a carrier of more verbose data now is IMHO a
wrong decision, and goes counter to the paved cow-path design pattern that
the WHAT WG holds so sacred and I believe causes as much pain as it does
gain.  Is it worth it? (Methinks no)

JF

Received on Wednesday, 29 April 2009 06:05:07 UTC