Issue 465: Proposal to address conformance through pointing device and voice input

Hello,

I'm curious how people react to this proposal to address
issue 465 [1]: conformance for keyboard and pointing device
and voice input.

The current conformance model is "subtractive": you have to do
everything for unconditional conformance. What you don't do you must
announce in your claim (in essence subtracting some requirements from
the full set).

The model used to be additive: there was a core set of requirements
and you added to them in the claim. But this was complicated and
necessitated the definition of a "core set" of checkpoints. I'm glad
we don't have the concept of a core set of checkpoints anymore.

In between (it appeared in no draft, I think), I tried a hybrid model
where the content labels were subtractive and the input modalities
were additive. I abandoned it because I thought it made life more
complicated. I'm not so sure, given recent discussions we've had on
this topic. 

Please consider the following proposal.

--------
PROPOSAL
--------

a) Checkpoint 1.1 only talks about the keyboard. Based on the
   9 March draft, that would give:

   "1.1 Ensure that the user can operate the user agent fully through
   keyboard input alone."
 
      "Note: For example, ensure that with keyboard input alone, the
      user can interact with enabled elements, select content,
      navigate viewports, configure the user agent, access
      documentation, install the user agent, etc.

b) Add the following to the Note (this is informative only):

    "This checkpoint means that all input device requirements of this
    document must be satisfied through the keyboard. It is also
    possible to conform for pointing device and voice input; see the
    section on input modality labels."

c) Do not add checkpoints for voice/pointing device input.

d) I'm not quite sure what to do with the Guideline itself.
   We resolved to delete the existing checkpoint 1.2 (since subsumed
   by 1.1). That leaves two checkpoints, one for keyboard input
   and one for text equivalent message in the UI. This means
   that the Guideline is not really about "input and output
   device-independence" any more (though that's a good principle!).

   I can't find a good place to put checkpoint 1.3 (i.e., it doesn't
   seem to fit in other Guidelines). 
   
   Suggestions welcome!
 
e) Do not change the section (3.1) on input modality labels.

f) Update the conformance model (3.1) since input modality
   requirements are no longer subtracted from the equation.

g) Change condition 2 of section 3.3 (well-formed conformance claims) 
   from:

      "Each input modality label is an assertion that the user agent
      does not satisfy the requirements associated with the label."

   to 
  
      "Each input modality label is an assertion that the user agent
      satisfies the requirements associated with the label."


----------
ADVANTAGES
----------

There are at least two advantages to this approach:

 a) I think it resolves Aaron's concern since the seeming
    contradiction is dropped.

 b) In general, it will make claims shorter since you won't have
    to say all the time "I don't do voice or pointing device input."

    Note: I have the *opposite* expectation for content type labels: I
    expect conforming user agents to satisfy requirements for most
    content type labels, and therefore there will be fewer exceptions 
    to document in a claim.

--------
QUESTION
--------

I hope people agree with me on the following point. Consider
checkpoint 12.3:

    "12.3 Document the default input configuration (e.g., the default
    keyboard bindings). [Priority 1]"

It is my understanding that when you claim conformance for the
keyboard only, that you only have to satisfy this checkpoint for the
keyboard. If you also claim conformance for pointing device and voice,
you must also satisfy this checkpoint for those input devices. If
you don't claim conformance for pointing device, you don't have to
document the default pointing device configuration.

If people agree with that (and the same reasoning applies to other
checkpoints, e.g., following operating environment conventions), then
I don't think there's a problem with the above proposal. 

If people don't agree, then we need to figure out why there would be
some partial requirements for some input devices even when there is no
conformance requirement for those devices.

Thank you,

 - Ian

[1] http://server.rehab.uiuc.edu/ua-issues/issues-linear-lc2.html#465
[2] http://www.w3.org/WAI/UA/WD-UAAG10-20010309/

-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel:                         +1 831 457-2842
Cell:                        +1 917 450-8783

Received on Saturday, 17 March 2001 12:18:19 UTC