more than two activation protocols to look at

[Sorry, I think one needs to read the whole thing]

At 08:59 PM 2002-01-05 , Charles McCathieNevile wrote:
>On Sat, 5 Jan 2002, Al Gilman wrote:
>  At 12:40 PM 2002-01-04 , Ian B. Jacobs wrote:
>[big snip]  >
>  >Please indicate what other questions WAI or the DOM WG or the
>  >HTML WG should answer.
>  Here are some of the related questions that appear to bear on the support
>  events in future versions of the DOM and XML-based languages (a.k.a.
>  modules, ...) and accessibility consequences of those features:
>  Q1.x Given the capabilities/constraints of current formats, what
>  information is
>  available which would be useful to users in making an informed decision as
>  activating or not activating a behavior which the author has set to be
>  triggered by a device-specific event?  An adaptive rebinding is required
>  either
>  when the UI adaptation of the user does not generate that event or does not
>  allow the un-confirmed automatic activation of the event handler for some
>  reason.  How would the available information be processed in a
>  adaptive interface binding (or bindings).  Some of this is guesswork or
>  "repair
>  text."  What are the options there, and the tradeoffs?
>CMN I don't see how, in the case where a user agent setup means that an event
>must be confirmed, it affects us. The event is either triggered or not, and i
>it is there is a confirmation interaction required, and if it is not then it
>is the simple case of a user agent configuration taht doesn't naturally fire
>that event.
>So I think we could simplify the cases in the question  to "where the user
>agent configuration does not generate that event".

Not quite.  There is a difference between the protocol where a dialog is
automatically triggered and where the user has to ask "what are my options for
action here" to expose the "well, there is some sort of a response to an
onChange applicable [or attached] here" information.  And the availability of
prompt information clarifying the likely outcomes of undertaking the action
will affect how the user uses this protocol-selection capability.

There is an existence proof for a "justDoIt, asMe, don'tEvenThinkAboutIt"
tri-mode setting which is currently available (scoped to all scripts) in IE,
for example.  In phase 2 under "what would you really like to give the user in
terms of tuning the delivery context" we can look into applying that kind of
protocol diversity to more narrowly-defined classes of response-actions.  I
classes of response actions because there are no non-semantic actions.  Only
actions that fall within different aspects of behavior.  Which aspects are
useful and/or risky, for which users, is something that is a user variable as
best as I can figure out, and not knowable from the content alone or by the
author a_priori (without a lot more HCI-for-PWD expertise than I have or want
to demand of authors).

We can expand on the "negotiating the shape of the table" topic more in what
follows.  Let me stop here just to say that while we can look at gradations in
action-taking protocol as a separate aspect of the solution, it is not enough
to think that there are only two protocol modes.  Even with the current
of information about the nature of the response to be expected.



Received on Tuesday, 8 January 2002 15:27:06 UTC