Re: W3C Web Accessibility Initiative Statement on Web Access Report from UK Disability Rights Commission

Tina Has it the nail on the head here.  At the risk of putting myself out of
a job, while it is true that automated tools cannot do it all (without a lot
of work), For testing against guidelines, you need only people who can test
well whether they are disabled or not.  If however, you are going for
usability and have criteria to bump against rather than some veague notion
that you need to test for usability, you will need some sort of sampling of
either people who understand the issues of the target audiences or from the
target audiences them selves.  I am often asked to look over a site by a
tester and comment on it re the guidelines since even if they test well,
What is percieved by someone who uses a different environment and or faces
different challenges may be different than what the tester percieves and we
need that perception if sites are to work.  People who use screen readers
have different spatial issues for instance than do people who may be testing
against the guidelines.  This does not mean that the tester will fail
without this perception, but the perception can characterize any need for
change that might constitute reporting for change to improve the site re the
guidelines if necessary.

----- Original Message ----- 
From: "Tina Holmboe" <tina@greytower.net>
To: <w3c-wai-ig@w3.org>
Sent: Wednesday, April 14, 2004 8:11 AM
Subject: Re: W3C Web Accessibility Initiative Statement on Web Access Report
from UK Disability Rights Commission



On 14 Apr, Julian Voelcker wrote:

> I forgot to mention, this was one of the points being stressed in the
> interview - you need to get sites checked by people with disabilities
> and cannot rely solely on the automatic systems.

  I've heard this argument before, and I still don't agree with it. I
  fail, you see, to understand where you'll get the *extraordinary*
  amount of people needed to cover all the various groups and
  accessibility issues.

  Mr. Brown, the old butcher who has just begun to use computers
  despite his old age and low vision, has quite different issues than
  Ms. Blue who, despite being blind, has grown up with 'puters of the
  latest mark all her life. Mrs. Wazi the local laywer from India with
  her university degree but a touch of twiching that makes using the
  mouse difficult has yet other issues than old Mr. Khaali who has
  perfect vision but isn't too good with English and really would like
  some help with all the abbreviations.

  In other words, an elderly immigrant with low vision have different
  problems in accessing information than has a young, blind, "native"
  citizen. This, for instance, can clearly be seen by the way some
  suggest that non-latin scripts should be implemented as graphical
  images because Unicode is "unreliable"[*].

  You also need more than one tester from each group so as to smooth out
  any variations due to personal preference and idiocyncrasies. You will
  also have to use the *same* testers for *every* website, in particular
  when testing a large number of sites in a certain area, say local
  government, so that you avoid having different testers having
  different personal preferences and different personal views - they are,
  after all, not testing compliance against guidelines, but how well the
  site works for *them* - not anyone else.

  Consider this. We don't maintain building codes. Instead, we test new
  architectural designs with disabled people. In short, we have a fellow
  in a wheelchair drive through the doors and see if they are wide
  enough. Grand, that. It was indeed wide enough . Then we build the place,
  and along comes old Mr. Smith in his trusty old, and specially built,
Permobil
  which happens to be about twice the width of the modern Chairman 2 Corpus
  used by the tester.

  This far we've not even *touched* upon the problem of whether to let a
  tester use his/her *own* setup/system which he/she is accustomed to,
  OR whether to let them use YOUR system which they ain't - and how
  *that* will impact the result ...

  The reason why we have, and agree upon, guidelines can in my view be
  summed up as this.

    * We cannot test everything using automated systems. It's impossible,
      even if people claim it isn't.

    * We cannot test everyone for everything. Still not possible, either
      in practical or economical terms.

    * So we do the best we can. We get together, we agree on guidelines,
      we ask those most impacted, and we make the guidelines both general
      and specific. Then we revise, take other issues into regard.

    * And finally we test against the guidelines using automated tools and
      people who are experts in testing. Whether they are disabled or not
      doesn't really matter; what *matters* is that they are able to be
      objective, perform the same tests time and time again[**], and that
      they are given solid guidelines.

  'tis not much different from any other area. You can't test-crash a new
car
  model with every conceivable size and shape human being; so we make
guidelines
  for the manufacturers to follow.

  It does also help if everyone and his grandmother would stop trying to
bloody
  re-interpret said guidelines to fit their own personal or political
agenda. But
  I won't hold my breath on that one.


 [*]
  NOT my advise, no. I'm still trying to find out which "programs" the users
  probably don't have after being told one should use images of, say,
Devangari text
  instead of Unicode.

 [**]
  The ability to survive boredom helps.

-- 
 -    Tina Holmboe                    Greytower Technologies
   tina@greytower.net                http://www.greytower.net/
   [+46] 0708 557 905

Received on Wednesday, 14 April 2004 09:41:29 UTC