Re: Checkpoint on testability

(Admin note: this has been Bcc'ed to AU and ER so they see the discussion,
but I think it is best to keep it followed up in GL. Anyone have a better
idea of how to do this so we keep the discussion moving at a reasonable pace
for everyone?)

summary:
This stuff makes good techniques. It may make for checkpoints/guidelines, and
it may belong under ATAG and WCAG differently.

Thought process:

It seems like a content requirement at some level to identify the
accessibility of content. (This is already in WCAG 1).

It seems that there is a lot of value for authoring tools in having that
identification in a way that can be tested adn understood by the tool. If I
open a page in tool X, and there is information that says "fragment #Y meets
the following requirements..." then editing that fragment may change that
status. Ideally, tool X would be able to verify the original statement, and
would know if it had changed as a result of editing.

(In fact the ATAG requires that information not be lost. But it allows for
non-reversible transformations. For example in converting a SMIL presentation
to some other format there may be no way to carry Audio descriptions in an
optional track - they would therefore have to be included as part of the main
track, although it may not be possible to seperate them out again
automatically.)

Clearly, using a verifiable method of ensuring some content is accessible is
of value to authoring tools. (Not least to satisfy the desire of most
designers for the least possible intrusion on the part of the tool...)

So, it seems simple to suggest that there are techniques which suggest both
the principle and the practise of making things accessible in a way that is
testable.

Going from there to checkpoints or guidelines requires answering the question
"is it always something that must be done, or should be done, or may be done
to provide access to content (or tools in the ATAG case)?" For example,
allowing configuration of prompting and alerting about accessibility problems
in content is not a checkpoint in ATAG. It is a generally recommended
technique, but there may be instances where someone develops a tool for a
specific type of editing, and doesn't allow the user to avoid being prompted
to fix some problem. Although this is hardly likely to be a mass-market
winner,  it doesn't prevent accessibility at any level, so this alone would
not be enough to make it fail any conformance requirements.

Cheers

Charles McCN

On Fri, 29 Dec 2000, Leonard R. Kasday wrote:

  Thanks all for the comments on the first draft of "checkpoint on testability".

  I can see now that I have to explain this guideline better.  I'll admit  I
  don't see it spelled out in the scope of the current WCAG charter, but I
  think its an important guideline that needs a home.  So I'll explain the
  guideline first and return later to the question of whether it belongs in
  GL.  I'm going to rephrase it a bit:

  Minimize the amount of human effort needed to test if content is
  accessible.  For example, if two techniques produce equally accessible
  content, pick the technique that requires the least human effort to test
  the result.

  Why? There are two reasons.

  The first reason is that in the real world, there's only finite time
  and  money available  for testing.  And in the long run at least, after
  we've automated all the testing we can, it's human effort that takes time
  and costs money.  So if a site takes too much human effort to test, it will
  not get tested completely, or might not get tested at all.

  Who does this affect?  it affects web authors who want to test their own
  work.  It affects system test departments in larger operations.  It affects
  organizations doing acceptance testing of web design work they have
  contracted out.  It affects organizations like schools who need to test if
  web resources are usable by all their students.   If testing takes too much
  effort for any of these organizations, it doesn't get done completely or at
  all.

  So ultimately it affects the end user.  Because in today's world, if
  something isn't being tested, it probably doesn't work well or at all.  So
  if these organizations haven't tested, what's passed on the end user is
  less likely to be fully accessible.

  And now for the second reason. This follows up a comment William made.  The
  end user to check for him or herself whether a site is accessible before
  trying to make use of it.  The more that testing can be automated, the more
  of that testing the user can do for him or herself.  That's because testing
  that takes human judgment costs the user time at best; and at worst it
  requires a type of judgment that the user can't do at all (e.g. a judgment
  requiring visual inspection of images for someone who's blind).

  Now, we can debate whether this issue is in the scope of GL, but I think
  its quite important nonetheless.

  So: is in in the scope of GL?  Well, on the one hand,  I don't see it
  spelled out in the charter.   But it is a requirement on content.  So what
  other group would do it?  As a practical matter, do we want to have two
  groups writing two different documents about content?

  Seems to me GL is the best home for this Guideline.

  Len
  --
  Leonard R. Kasday, Ph.D.
  Institute on Disabilities/UAP and Dept. of Electrical Engineering at Temple
  University
  (215) 204-2247 (voice)                 (800) 750-7428 (TTY)
  http://astro.temple.edu/~kasday         mailto:kasday@acm.org

  Chair, W3C Web Accessibility Initiative Evaluation and Repair Tools Group
  http://www.w3.org/WAI/ER/IG/

  The WAVE web page accessibility evaluation assistant:
  http://www.temple.edu/inst_disabilities/piat/wave/


-- 
Charles McCathieNevile    mailto:charles@w3.org    phone: +61 (0) 409 134 136
W3C Web Accessibility Initiative                      http://www.w3.org/WAI
Location: I-cubed, 110 Victoria Street, Carlton VIC 3053, Australia
until 6 January 2001 at:
W3C INRIA, 2004 Route des Lucioles, BP 93, 06902 Sophia Antipolis Cedex, France

Received on Sunday, 31 December 2000 10:50:42 UTC