Re: usability/accessibility. Yes; examples.

The approach that Al is describing is in fact what the Authoring Tools group
is working very hard on at the moment.

The basic idea is that the working group are probably the best qualified
people to make assertions of "I/we thnk that technique X meets the
requirements of checkpoint Y", and that if we do provide examples of such
assertions, and discussion of them where it arises, then people who have not
lived and breathed the document through its genesis will be more comfortable
in using it to make similar assertions of their own.

One of the things we have found is that companies are reluctant to make an
assessment of their work when they don't have much more than the
specification to go by - test suites and example assessments appear to be
very important to them.

(The rest of this post is a lengthy attempted explanation of how the above
is not specific to us, and has been done in other areas for centuries. Feel
free to ignore it)

Although these recommendations are not legal requirements (unless someone
passes a law to make them so, and in that case it is their responsibility to
work out how to administer that law) there is a paralell with how law works
in many countries...

Parliament (in Australia) passes a law. Often this has been drafted by a
committee of people with expertise in the area, tossed around for public
discussion to get some more expertise applied, and discussed in parliament
itself, but that is not required - it is possible in principle forone member
to present a bill, and get a vote on it without discussion.

The law just says "Thou shalt do such and such, or something will
happen". Let us use as an example a law requiring people to grow their hair
long (I believe that used to be a real law in some places).

Let us imagine, further, that a policeman decides that I have not done what
is required, and brings a case against me. 

(Diversion: In Australia, the USA, and many other places there are two kinds
of law: civil and criminal. In civil law, some private person applies to the
courst to try a case in which they allege that some other person has broken
the law in a way which affects them - for example parking a bus in their
living room, and so spoiling a dinner party. The court examines the claim,
decides if there was something illegal done, and what should be done to
redress it. In Australia they can also decide who should pay for the process.
In criminal law, "the Crown" (in the US this is normally "the People") as
represented by the Police and various other officers appointed on behalf of
the Queen or government brings notice of a breach of the law to a court. The
court then decides if the law was breached and what should be done about it)

If I am the first person charged under the law, then the court may be the
place wherevarious questions are answered, such as "how long is long", and
whether it means my beard - things which are deemed relevant, but are not
specified in the bill that was passed. These decisions, and what motivates
them, are generally recorded. Often experts in the area are called in to help
the court make the decisions.

These decisions then become precedents. Parliament may make further
regulations, or change the legislation, in response to how it is being
applied, (or for any other reason it sees fit), but where there is not an
answer in the legislation to a question that has been decided in a court,
subsequent cases are obliged to recognise that the question has arisen before
and been answered, and to be consistent with the previous answer or justify
why they are not.

So the imnportant thing here is that we have specifications that have been
written and agreed to by "the Consortium" (our Recommendations) as good for
the Web, and we have people who are capable of understanding what was
intended by the Recommendation and how it applies to specific examples (the
Working group). How the experts (the Working Group) interpret the
Recommendation is useful as further help or an example for other people
dealing with similar specific questions.

(The rest of the example falls down - we don't have any authority over what
people do, just what we say about it. And in tricky cases we may not all
agree, and we may decide that there was something that we need to fix because
the Recommendation didn't deal with some particular case as we would like it
to.)

Charles McCN

On Mon, 3 Jul 2000, Al Gilman wrote:

  By George, he's Got It!
  
  At 08:42 AM 2000-07-03 -0700, William Loughborough wrote:
  
  >What I gather from this is that we must attend more to the
  >"accessibility of the accessibility recommendations" rather than staying
  >focused almost entirely on the recommendations themselves.
  >
  [...]
  
  And just in case I haven't said it 40 times yet, that means examples,
  examples, examples.
  
  The legal connotations of the term 'certification' are an obstacle, for
  sure.  Let the W3C avoid certification, indeed.  But just as we adopted
  'recommendation' as a euphemism for under-the-radar effective standards,
  why doesn't the WAI recognize it _is_ entirely appropriate for the W3C to
  get involved in the 'authentication' of examples.  Designers won't know
  what to do without examples.  
  
  Recognition (who said endorsement -- It's not an endorsement!) from the W3C
  is something that those who invest in compliance early will value.  And it
  has a ripple effect.  It is good for our purposes of rapid adoption if we
  can do anything good for those who listen well and early.
  
  It is appropriate, and highly beneficial for training results, for the W3C
  to make assertions of the form "[This resource] is an example which
  satisfies WCAG 1.0, checkpoint n.m."  The manual checks that Bobby leaves
  the person to do are not 100% reliable yet on a self-administered basis.
  Identifying real examples where WAI insiders have confirmed the assertion
  of compliance _adds significantly to the clarity with which the rules are
  understood, and the speed with which they will be adopted_.
  
  Cross-tabulate the feature utilization at the EASI winner sites with
  checkpoints that bear on these feature-use examples.  Double-check if the
  use is a good example.  Publish this database.  Add SNOW to cover some
  features or techniques not well illustrated in the EASI batch.  SNOW is
  rich with very consciouly applied techniques done up to life size.  
  
  This will complete an initial critical mass of technique examples.  This
  will draw people saying "Yes, but I have another technique that works
  well."  From there it takes off.
  
  Al
  
  
  

--
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
Postal: GPO Box 2476V, Melbourne 3001,  Australia

Received on Tuesday, 4 July 2000 03:10:36 UTC