Chuck Hitchcock comments on last call UAGL [Fwd: User Agent Guidelines 1.0 Comments]

Hello,

I am forwarding comments from Chuck Hitchcock on the last
call User Agent Guidelines [1]. My comments are intermixed.

[1] http://www.w3.org/TR/1999/WD-WAI-USERAGENT-19991105/

Chuck Hitchcock wrote:
> 
> Ian and Jon,
> 
> All in all, I think that the UA Guidelines are terrific.  A few comments
> follow.  This includes contributions from Michael Cooper and is offered
> without the benefit of having reviewed recent UA listserv messages or
> meeting minutes.
> 
> Please feel free to use what helps and ignore the rest.

Thank you for the review, Chuck!

> 
> In the Introduction in the section related to user interface, it might be
> helpful to include a comment about providing a "reset defaults" function for
> user settings.  This can be global or by groups of settings but I find that
> many individuals fuss with settings to a point where they forget the
> original recommended defaults and really need to start over.

This sounds like a Technique, but the WG should consider whether
this should be a checkpoint, say in the section on configuration.
 
> In Checkpoint 1.1., Principles of Accessible Design:  I am not aware of any
> real evidence indicating that contextual access through cascading menus is
> supportive of those with cognitive impairments although common sense would
> seem to indicate that this might be true.  On the other hand, menus that are
> not stable or predictible might also be problematic for some.  The
> organization and grouping of menu items is much more important.  I would
> change the language in Contextual access to read:
> 
> -->Contextual access (e.g. through cascading menus, through help systems,
> etc) may help users with cognitive impairments and others unfamiliar with
> the tool.

Unless there are objections, I will use your language.
 
> Checkpoint 1.4: Wonder about the use of the word "every". The explanation of
> the checkpoint mentions that it should be efficient enough to support
> production use - which I think is appropriate, it can't be possible to do
> everything through the keyboard or we wouldn't need other forms of input.
> But the "every" in the first sentence contradicts that.

I would argue that if you can't do something at all with a given device,
the UA is broken. I can enter text with the mouse if I have an onscreen
keyboard. However, I don't want to do so since I can use the keyboard
(and for that particular functionality, I don't want to require the
UA to support text entry with the mouse; it's still conceivable,
however).

Thus, in my opinion, users who can use different devices should be
able to use the one best-suited to their abilities and the task at
hand. Users who can only use one device should be able to operate the
tool with that device alone.
 
> Checkpoint 2.1: Seems a little vague. "Alternative content" is defined above
> but not normatively; in the absence of a definition here (which is probably
> done to avoid the mistake of leaving something out), it would be difficult
> for anybody to claim full conformance.

Yes, I think we need to add a definition.
 
> Checkpoint 3.9: The word used in WCAG is "redirects" instead of forwards and
> I think it should be the same here to reduce confusion.

This may have been a conscious change since the WCAG, but I don't
recall. Comments from others?
 
> Typo in "Note" just before 4.1 ("in including").

Ok. (Fixed).
 
> Guideline 4: There should be a technique that says "Allow user to define and
> apply in the cascade any number of custom style sheets." There might also
> should be a P3 new technique "Allow user to define custom styles for found
> class attribute values in the document".

I'll add to the techniques document.
 
> Checkpoints 4.14, 4.15 and 4.16: Should there be a qualifier here to
> indicate that the speech controls apply only to browsers that include native
> synthesized speech?

I don't know that this is required since the conformance clause
says:

       "User agents must satisfy natively all the applicable 
        checkpoints for a chosen conformance level."

Thus, browsers that do not synthesize speech natively don't have
to satisfy the checkpoints.

>   I suspect that unless one is using a custom talking
> browser, a shell program that uses component browser technology and speech,
> or a screen reader, the speech controls may not apply and may cause
> confusion during conformance ratings.  As an example, in CAST's eReader, we
> use the IE component and the SAPI style controls.  Those speech controls are
> not part of the browser but are part of the shell application within which
> the browser with speech & highlighting functions.
> 
> Checkpoint 4.17: The "note" does not make sense. In a stylesheet-based
> browser, there must always be at least one stylesheet, otherwise the browser
> has no way to know how it should render any content. In other words, it
> should not be possible to turn off the user agent default stylesheet, just
> to override it with a custom one.

Agreed. Refer to issue 132
http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#132
 
> Checkpoints 5.3, 5.4, 5.5, 5.6, 5.7 seem to be special cases of 5.1 and
> should be identified as such.

Checkpoint 5.1 needs clarification, which may have an impact
on this situation. Refer to issue 125.
http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#125
 
> The preamble to Guideline 8 mentions separately scrolling table regions
> identified by THEAD, TBODY, and TFOOT. But there isn't a checkpoint
> requiring user agents to implement that. Don't know if that's expected to be
> covered by HTML 4 conformance but it might be good to mention here.

You're right. The reference to table regions seems out of place
here. It might be better closer to 7.3 (if 7.3 survives).

I think it should be removed from the rationale and either
moved to G7 or to the techniques.
 
> Guideline 8:  In the paragraph that follows the 3 bullet items, I would
> change "learning disabilities" to "cognitive disabilities".  This is a bit
> broader and will include those who have orientation problems that are not
> related to specific learning disabilities.

Refer to issue 137.
http://cmos-eng.rehab.uiuc.edu/ua-issues/issues-linear.html#137
 
> Guideline 9:  Also suggest changing reference to "learning disabilities" to
> "cognitive disabilities" in the first paragraph.

Same comment.

Thanks again Chuck,

 - Ian

Received on Wednesday, 1 December 1999 14:12:35 UTC