--
_______________________________________________________________
Thierry MICHEL tmichel@w3.org
W3C / INRIA +33 (0) 4 92 38 79 87
2004, Route des Lucioles
BP 93
06902 Sophia Antipolis Cedex,
France
_______________________________________________________________
Forwarded message 1
Most of these comments are to do with improving the accessibilty of the
specification itself, or of the examples that it uses:
The last example in section 3.2.1 demonstrates an animation being raised 5
seconds after a click event. Since we hope not to have click events, we
should ask that this be changed to an activate event. WCAG checkpoint 6.2
requires device-independence for input triggers.
Same comment arises in relation to most of the examples in 3.3.2
It is noted that some more pictures would help - indeed they would.
in 3.3.5 an example is given that is based on a mouseover. Although the
example appears to be inherently spatial, removing the device-dependence (and
replacing it with a focus event, or perhaps in this case a select or
activate) would provide the input dependence required. Since the rest of the
example is hypothetical it doesn't need to be taken further, although a model
of how this might work would allow for either selection, or a means of
navigating independently rather than purely serially.
3.3.6 again has a number of click examples - more cases for a simple
activate.
Section 3.5 describes the sandwich model, noting that changes to CSS rules
will still be subject to the CSS cascade, which is an accessibility
requirement. It is probably worthwhile at that point making a brief statement
that another effect of this is that animations may not in fact be rendered,
so important changes to content must not be specified purely by animation of
style (as per WCAG 6.3). This is implicit in the final pragraphs of the
section, but making it more explicit would be helpful.
Figure 4 - the state diagram for start, stop, restart, freeze, would be a lot
better with a short description in a longdesc, d-link, or immediately below
the image. Alternatively, the alt should say that the diagram is explained in
the following section.
3.7.3 again uses click as an example of a user input event.
The example at the end of 3.7.4 is click-driven again, and uses the link
content "Click Here!" (contrary to WCAG checkpoint 13.1)
3.7.5 starts with another click-based example
The final example of 4.2 uses mouseover / mouseout, and would be better off
using focusin / focusout. Whether a corresponding adjustment in the link text
is strictly necessary is a moot point - one could expect users of assistive
technologies and mouseless systems to go the extra distance in learning to
use their interfaces and not need the behaviour explicitly described,
although it would be nice if the talk was device-independent as well as the
walk.
The example of illegal SMIL in 5.4 should nonetheless provide alternative
content for the image - an alt attribute at least.
Other than that we think the question of giving assistive technologies access
to the animation effects is a (soon to be fairy important) matter for User
Agent Guidelines.