W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2001

upside AND downside and readily achievable measures

From: Al Gilman <asgilman@iamdigex.net>
Date: Sat, 27 Oct 2001 12:20:39 -0400
Message-Id: <200110271609.MAA1065127@smtp1.mail.iamworld.net>
To: w3c-wai-gl@w3.org

In the consideration of what readily implementable measures authors should
be expected to implement, it is not necessarily appropriate to limit
ourselves to damage control.

One thing we are dancing with in the NCITS V2 work is a possible test that
[the equivalent of authors] should take all measures that are readily
achievable _and improve the upside returns to the high end_ of feasible
user-side interventions.  Not just to take measures that curtail the
downside, that eliminate predictable total failures, but that enable
substantial gains in usability through corresponding measures taken on the
user side, where the gains would not be feasible if the [authors] failed to
take these readily achievable measures.

Admittedly in that debate, and in consideration of "device independence
principles" as drafted by the DI group, I would argue that there are [close
enough to always] some identifiable people with disability whose _failure
threshold_ depends on these higher levels of user side intervention, that
are only widely implemented in job accomodation situations.  But if we make
the upside test what is theoretically feasible for the most multiple and
severe condition scenario that we can push the theoretical cases to, it may
not be viewed as reasonable by the [website author and mass market product]
target audience of the rules.

Job accomodation with speech command has a "best current practice" pattern
of requiring the development of a macro command library so that the user
has shortcuts for tasks commonly performed in that job that offset the
speed disadvantage of speech command with the as-shipped vocabulary of
application-control commands.

Maybe we should state a test that author measures which are a) readily
achievable and b) necessary for workgroup-accomodation-reasonable levels of
user-side intervention are also required, in addition to those that are
required to eliminate outright failures.  

In the workplace context, elimination of outright failure to access the
information is often not enough to make the person employable within the
workflow of the organization.  In the workplace, the person with a
disability has to be able to fit into the general flow of work through the
workgroup with not more than a reasonable dent in the group throughput.
This often means that the information access has to reach a usability level
above the "useless for all practical purposes" threshold that I nominated
as our 'failure' test for the minimum set analysis.

Websites exposed on the public Internet are like mass-market products
placed on the open market in that there is [at least precedent if not]
reason to argue that any readily achievable measures should be taken
because of the leverage effect and the fact that counting on end-users to
do something is generally not efficacious.

Definition:  distinguish that an intervention can be shown to be
_effective_ under controlled laboratory conditions, but to be viewed as
_efficacious_ it must be demonstrated to generate the intended outcomes
under field conditions, under the proposed ecological conditions of
dissemination and use.

This distinction in terms comes from the health care policy community.

Received on Saturday, 27 October 2001 12:08:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:16 GMT