Re: Working Group Decision on ISSUE-31 / ISSUE-80 validation survey

On Thu, Apr 21, 2011 at 6:54 PM, John Foliot <jfoliot@stanford.edu> wrote:
> Do content owners and developers
> think that meeting an arbitrary validation check-point results in the best
> user experience for their users?

In many cases, yes.

> If yes, why[?]

Because some customers and users of web software judge it partly based
on whether it validates.  If software's HTML output doesn't validate,
some percentage of users will hold that against it, and assume that it
reflects some deficiency in the software (whether or not they actually
understand the errors).  At least in the open-source world, there also
tends to be social pressure among developers to conform to standards.

> Simply put, what is more important, a validation green-light badge, or an
> inclusive user experience?

In my experience as a web developer, only a tiny percentage of users
of typical web applications use anything other than conventional
browsers (desktop or mobile).  For typical commercial products,
there's essentially no incentive to include disabled users in testing,
planning, or QA, because the expense of accommodating them exceeds any
likely RoI.  For non-commercial products like MediaWiki, developers
are often volunteers who are mostly interested in their personal use
of the software, or use by the users they come into contact with (who
are overwhelmingly not disabled), and again they have little incentive
to make more than the most basic accommodations.  In all cases,
developers would probably be willing to make simple changes if they
thought it would help disabled users, but only if they didn't require
a lot of resources or effort.

For instance, as a MediaWiki developer, I could change the default alt
text produced for user-provided images from the empty string to the
filename or something else anytime I felt like it in about five
minutes, and the change would most likely take effect on Wikipedia
within a few months.  But I simply have not been approached by actual
blind users who have requested such a change, to the best of my
recollection.  So I don't know if making a change would actually be
useful, or if so, what change I should make.

The current HTML5 spec says we should either use an empty string or
just omit the alt attribute, but I have no reason to believe that was
based on actual analysis of screenreaders in practice rather than
theoretical principles.  I took a brief look at WCAG 2.0 just now, but
it seems to just completely ignore my use-case (user-uploaded image,
no idea how it's being used).  I looked at "HTML5: Techniques for
providing useful text alternatives", but it just says I have to
provide a caption instead, again ignoring my use-case.

So I'm left with only one standard that even addresses my (extremely
common) situation, and we appear to currently follow it.  If I really
cared, I could go out of my way to track down blind MediaWiki users
and ask them what the best behavior would be, maybe providing them
with mockup pages.  But I have lots of other things to do, and I've
received no actual user complaints, so I'm not going to.  It just is
not worth the effort for me.

Received on Thursday, 21 April 2011 23:20:32 UTC