W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2003

My comments on WCAG 2.0 proposed draft

From: Arditi, Aries <aarditi@lighthouse.org>
Date: Tue, 09 Dec 2003 14:13:45 -0500
Message-Id: <5.2.0.9.2.20031209141338.020d5980@localhost>
To: w3c-wai-gl@w3.org

Hi all,

Wendy Chisholm and Judy Brewer asked me to comment on this months ago, and 
I posted it to the site they asked me to post it at.  Recently, someone 
more savvy in the functioning of the WAI lists informed me that the list is 
totally ignored.  So here, I repost it, in the hopes that it might be 
helpful to someone, somewhere, some day....

Aries Arditi
Senior Fellow in Vision Science
Lighthouse International

Original Post:

With respect to the June 24 WCAG 2.0 draft, I have a number of comments,
detailed below.  First, however, I would like to congratulate the authors
for taking on such a difficult task.  In particular, I'm delighted that the
group is intending to broaden the scope of the guidelines to address a
broader set of technologies.  To the extent that this broadening succeeds,
its influence can go well beyond that of web content and authoring, and
potentially guide the efforts of developers of assistive technology as well.


Having said that, I believe also that the guidelines are in general less
comprehensible (indeed, less accessible!) to the intended readership than
were the version 1.0 guidelines, and that the lack of comprehensibility
stems from the very abstraction of principles that was necessary for
broadening to other technologies.  I believe these problems can be addressed
and alleviated through the planned technology-specific checklists,
especially if the checkpoints in the WCAG are cross referenced to technology
specific guidelines. Apart from a few exceptions, I don't believe, however,
that making the guidelines significantly easier to understand can be
accomplished by simple rewriting.  Like a mathematics text in which axioms
and theorems provide the foundation, and examples and problems the practical
understanding, this may be a necessarily abstract document.

The conformance structure of Core and Extended is fine, both easy to
understand and potentially effective, provided more objective and
quantitative methods of evaluating specific conformance to checkpoints can
be developed.  But conformance to specific checkpoints is best evaluated in
the context of specific technologies, where specific items are less open to
braod interpretation.  Without technology specific checkpoints, I would
imagine that squabbling and bickering over conformance will increase, and
with more latitude to interpret the guidelines as they see fit, developers
could claim more for less actual conformance.  So I would favor keeping the
general structure of the draft as a kind of umbrella document, but
instantiate the conformance checkpoints with technology-specific
checkpoints.  Since the planned technology-specific checklists were not
included in this draft, it's difficult to evaluate the conformance
structure.  However, in theory, the structure seems fine to me.

Developers at my organization do use the 1.0 WCAG, but much of their use is
indirect---via automated tools to evaluate conformance.  Because web sites
can be huge and dynamic, reliance on such tools is at least partially
inevitable.  However, to the extent that the current document is even less
specific than v1.0, I would see migration being difficult. Without
technology-specific checklists, it is easy to imagine developers of
evaluation and repair tools with widely disparate views of what constitutes
conformance.  Again, I would prefer to view this draft as an umbrella
document with more specifics to be incorporated as an essential component of
the 2.0 WCAG recommendation.

Here are some other specific comments, nearly all pertaining to Guidelines 1
and 2:

1. I didn't understand the designation of examples and benefits as
"non-normative"  or "informative."  Are these terms lingo, or am I just
being thick (or both)?

2. Overview of Design Principles, last para before Note:

This would better read, "Accessible web content can benefit people with as
well as without disabilities." Accessible web content does not ALWAYS
benefit people other than the disabled. High levels of magnification, for
example, which can make things more accessible to people with low vision,
can make documents more difficult to navigate.

Delete next 2 sentences, which are superfluous since examples of accessible
web content benefiting others is subsequently given (and giving wheelchair
examples to the reader is tiresome---readers of the WCAG can, nearly 14
years after the ADA, be assumed to understand general things about
disabilities without being prompted to conjure up images of wheelchairs).

3. User needs, first bullet:

Change "via sound" to "visually."

4. Definitions for checkpoint 1.1: text equivalent

This is not a definition.  I would first define "text" as code representing
written language, that is a one-to-one mapping of alphabetic and numeric
symbols.  Then define "text-equivalent" as text that serves to communicate
substantially equivalent content as another representation such as an image.

5. Benefits of checkpoint 1.1, first bullet:

add ", or otherwise transformed to different presentation format (e.g. font,
text size)

6. Checkpoint 1.5

I believe that the benefit of adding structure to the document is that it
allows assistive technologies more information about the document in order
to tailor its presentation to the user's capabilities.  This checkpoint, as
written, however, seems to focus on how document elements are rendered.  For
example, the (sole) required success criterion refers to visual appearance
or auditory characteristic.  The examples also are about how to render.
That's not a WCAG issue, in my view.

Also, who (or what) could judge whether structure "has been made perceivable
to more people..."?  Compared to what? This is way too vague.

7. Editorial Note (22 May 2003) regarding "additional items in 4/29 draft

Yes, in my opinion, these should be moved to techniques.

8.  Checkpoint 1.6

This is very vaguely worded.  I would propose instead the following
language:  "Where possible, use layered designs to allow selective
presentation or blanking of visual or auditory elements, comprising, for
example, foreground and background content."

The sole success criterion for 1.6 is in terms only of visually imagery.  I
would refer to layered auditory images, such as background music with a
voice-over, to allow hearing-impaired user to remove the background music
for greater clarity.

Under best practice for 1.6 #1 and #2, what does "easily readable" mean?
And to whom? This is not quantifiable.

Also, who came up with the 20 dB (or 4 X louder) foreground/background audio
figure?  Are there any data to back this up as being sufficient?

It's true that there are no effective tests for visibility or readability
that do not rely on standard human observers in terms of color or color
contrast (the claims of some notwitstanding). The reasons for this include
the fact that display devices and hardware used to drive them vary widely in
peak luminance, color gamut, and linearity; viewing conditions of observers
vary widely (ambient room lighting, presence of reflections from nearby
windows, etc.); readability and visual discriminability also depend
critically on the kinds of stimuli used.  A very low contrast text in a huge
font may be as or more readable than a smaller font in very high contrast.
Typography and image characteristics also go into the mix of important
factors.  And let us not leave out the enormous variability in visual
capacity of the visually-impaired population, some of whom will settle for
no less than the blackest black against the whitest white. The long and
short of it is that we simply do not have in the foreseeable future a good
way to evaluate objectively things like accessibility of text contrast
and/or color.  This is thus an issue in which, I believe, the best we can do
(and thus the most we should do) is to suggest principles to follow to
enhance accessibility.

9. Examples of checkpoint 1.6 first bullet:

What are the "standard foreground/background contrast requirements"? If
there are any, please clarify (and cite).

10. Required success criteria for checkpoint 2.2 #1:

I would add the following bullet: "* or the user can control the
presentation by eliciting phases sequentially via keystrokes."

11. Checkpoint 2.3

I would be very surprised if anyone could devise a substantive and effective
test for flicker in machine readable format.  Here are some thought
experiments:

Consider a single pixel flickering in the 3-49Hz range on your screen right
now.  Do you think you could detect it?  If so, do you think it could
possibly cause a seizure in someone with photosensitive epilepsy?  You do?
How about if that single pixel flickered by changing its intensity 1/256? or
1/(256^3)?
Still think so?

How many phases (i.e. alternations) of light and dark constitute flicker in
the 3-49 Hz range? What if I present a black dot for 1.5 seconds, then
replace it by a white dot for 1.5 seconds?  Would you call that 3 Hz
flicker?  In other words, how many "cycles" of the pattern do you require
before you call it flicker?

Here's a third: Mathematically the sum of leftward moving and rightword
sinusoidal gratings is equal to a single stationary sinusoidal grating whose
bars reverse in contrast.  Any single point in that pattern will flicker. Do
we call that flicker or moving gratings?  Were we to analyze many moving
patterns we would find that we could describe them mathematically in terms
of flickering components. What if we have a video taken from a moving car or
train that passes by a picket fence.  Will not there be some flicker in that
video? Shall we ban all animations in the WCAG?

It is likely that scientists will some day be able to better zero in on the
specific visual causes of photosensitive epileptic seizures; presumably high
contrast large field flicker poses a greater risk than do low contrast
flickering single pixels.

Until we know more, or until qualified experts on this condition weigh in,
or credible scientific evidence is presented that better specifies the
parameters of the flicker, I believe this checkpoint should be deleted.

Thank you for the opportunity to comment on this draft.

Aries Arditi, Ph.D.
Senior Fellow in Vision Science
Arlene R. Gordon Research Institute
Lighthouse International
111 East 59th Street
New York, NY 10022

Tel: +1 212 821 9500 (direct)
Fax: +1 212 751 9667
<http://www.lighthouse.org/research_staff_arditi.htm>http://www.lighthouse.org/research_staff_arditi.htm


**************************************************************************************************
The contents of this email and any attachments are confidential.
It is intended for the named recipient(s) only.
If you have received this email in error please notify the system manager 
or  the
sender immediately and do not disclose the contents to anyone or make copies.

** eSafe scanned this email for viruses, vandals and malicious content **
**************************************************************************************************
Received on Tuesday, 9 December 2003 14:15:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:26 GMT