Re: Priorities & Impacts; Affected Groups

I would like to respond to comments made on 12/14/98 by Charles 
McCathieNevile (hereafter "CMc"). Note that "EH" indicates the responses of 
Eric Hansen.

CMc:
I am personally not keen on the idea of including elaborate impact
statements, although there are a number in the guidelines. 
EH:
I don't think that impact statements in the document would necessarily need 
to be more elaborate than what is included now. (Perhaps just a few 
refinements might be enough.) Currently the document highlights the _major_ 
affected disability groups and, in some cases, non-disability-related 
groups. I see the majority of impact data as non-normative and found in 
separate documents. 

CMc:
I am dead against the idea of quantifying them, as I feel that is an 
exercise which
is certainly beyond the scope, and probably beyond the resources of the
WG.
EH:
I see some kind of quantification as valuable for ensuring the documents' 
audiences that the circumstances of key disability groups and been examined 
in the context of each guideline. I have generated draft impact ratings for 
13 disability groups for 2 of the 17 guidelines. 

CMc:
How much impact a particular problem has on the general community (as some
kind of weighted average of impacts on subsections) is, in my opinion
irrelevant to the goals of the group. It does not assist us to determine
what guideines should be followed, and what techniques are available, to
ensure that web content is accessible to all. 
EH:
I think that impact ratings are one fairly stable piece of information that 
would go into figuring out which techniques should be followed. 
Specifically, impact ratings would be some of the information from which 
priority (importance) ratings would be derived. Furthermore, the impact 
ratings have the benefit of being much more focused than the priorities. 
(See earlier memos.)

CMc:
 It is also important in such a situation to have
accurate, up to date information. A W3C recommendation cannot be changed -
it is by definition a stable document.  [New para.] In this context, the 
removal of priorities on guidelines may be helpful. The guidelines 
themselves must be followed, but it may happen in some
months that there is no necessity for most people to do anything
particular to follow a given guideline, since the problem is handled for
them by an authoring tool (as an abstract example).
EH: 
Since earlier memos I have supported the idea that priorities don't belong 
in the normative portion of the document because they are too changeable. 
One of the major motivations for proposing the idea of "impacts" is to find 
a more stable indicator that might be included in the normative part (as 
well as providing a firmer conceptual anchoring of the guidelines to needs 
of specific disability groups). 

CMc:
The priority definitions are information which may be useful to people
trying to do a cost/benefit analysis - things which are priority 2
techniques are not going to provide access in themselves, they are just
going to significantly improve the quality of that access (by bringing it
nearer the quality experienced by 'mainstream users' - perhaps we should
make that more explicit in the P2 definition?) while things which are
priority 1 deal with the ability to access the information or function in
any form, without minimal regard to the difficulty of using that
information or function. 
EH:
I think that it is very important to make explicit how the priorities are 
derived. 

If I understand the nature of priorities, I must disagree with your 
characterization of "priorities" as _input_ to the calculation of 
cost-benefit. My understanding has been that priorities are actually the 
_output_ from a calculation (albeit informal) that uses cost-benefit types 
of considerations (cost, feasibility, practicality, availability of tools, 
etc.) as input. How can one generate an indicator of the priority (or 
importance) related to a technique without consideration of the cost and 
availability of the tools for carrying out that technique? Indeed, it is 
such considerations that make priorities changeable.

CMc:
(It could be argued that giving a person access
to a binary representation of the data in a GIF file is providing access,
but I think most people would agree that it is a stupid argument.)
EH:
That particular example might easily be dismissed, but I will propose 
another one that is more realistic. 

Certain individuals who are deaf or are dyslexic can perceive the written 
text, but the text is just as inaccessible to them as if they cannot 
perceive it (if I am exaggerating, someone please correct me). A definition 
of priorities that assigns priority 1 to issues of perception and priority 
2 to issues of language may give too little consideration to 
language-related issues even though they can be barriers that are just as 
severe.

One may respond, "Yet issues of perception are more fundamental than issues 
of understanding or comprehension. Issues of perception deserve to have a 
higher priority than issues of understanding and comprehension."

Yet, we should be aware that it is not very hard to recast issues of 
language as issues of perception. To illustrate this, I will propose a new 
technique or guideline: "If the written text is inaccessible, you must 
provide an alternative accessible version of the text." 

In other words, I as an individual with a severe print-related disability 
may say, "This technique is really no less important than the priority 1 
technique of providing alternative text for graphics. This technique should 
be priority 1, not priority 2. The alternative, accessible text must be 
made perceptible to my senses." 

By the way, this proposed technique is not that "far out." This technique 
or guideline encompasses the material in A.7, at least some of the material 
in B.3, part of A.5, and so on.


=============================
Eric G. Hansen, Ph.D.
Development Scientist
Educational Testing Service
ETS 12-R
Rosedale Road
Princeton, NJ 08541
(W) 609-734-5615
(Fax) 609-734-1090
E-mail: ehansen@ets.org

Received on Monday, 14 December 1998 18:21:32 UTC