Some comments on the user agent guidelines

The following is my input on the "last call" working draft of the
W3C's user agent accessibility guidelines:

*    In the guideline that mentions "relative position" as a
status message, also include "absolute size."  An example of its
use is being able to determine the size of the current web page
in a browser in order to decide whether to read it now or
download it for later reading.  The size should be stated in
commonly understood units of measure, sometimes in more than one
way to ensure it is readily understood.  For example, the playing
time of an audio file could be stated in terms of hours, minutes,
and seconds.  The size of a primarily text based web page might
be stated in both kilobytes and screens, where a screen of
information is calculated based on the current viewport (often a
window where the PageDown or Spacebar key invokes the next
screen).

*    When applicable (such as with a computer based browser),
provide a mechanism for downloading a batch of links to the
current page.  It is common for a user to want several linked
pages downloaded, yet most browsers today require that each link
be visited and separately downloaded.  An example would be a
dialog that includes a listbox of file extensions representing
files linked to the current web page.  The user could select one,
multiple, or all of the extensions to be downloaded in an
automated process.  The file names would be the same or closest
equivalents on the user's computer in a directory/folder that
could be chosen in another control of the dialog.  An enhanced
option would place all downloaded files in a compressed archive
using an industry standard format (such as the public domain .zip
one).

*    Support the resumption of partial downloads so that a time
consuming but incomplete download is continued rather than
restarted.

*    When the system focus is on text (either plain or enriched),
support navigation by character, word, line, paragraph, and
screen.  Precise and flexible navigation of this kind with a
system cursor is often useful when browsing with accessibility
aids.  Evidence of this point is that the leading Windows screen
readers have super-imposed such navigation on a popular web
browser that does not natively support it (e.g., Winvision,
Window-Eyes, and JAWS with Internet Explorer).

*    In the guideline that mentions searching, also include ways
of indicating the beginning and direction of the search (forward
or backward starting from the current focus, beginning, or end of
the document).

*    In the guideline about status messages, also include an
indication when a web page is finished loading, an audio file is
finished playing, a video file is finished displaying, etc.  The
mode of the "finished" status message should include text, audio,
and/or visual means of presentation.

*    In a guideline about keyboard input, also state that each
button on toolbars should have an equivalent keyboard shortcut. 
Mouse clicking such buttons is often the primary way that users
without disabilities perform common commands.  Given the benefit
of keyboard shortcuts for various disabilities and accessibility
aids, the user agent should be designed with complete equivalence
between buttonbar and keyboard operation.

*    In the guideline about documentation in HTML, also specify
that a manual should be available as a single, complete HTML file
when a system of multiple, linked files is provided as well.

*    In a guideline about viewports, also include a configuration
setting by which the application automatically maximizes the
current view port.  For example, the parent window of the browser
would automatically be maximized when launched, and each child
window would automatically be maximized when it received input
focus.  Maximizing does not necessarily mean occupying the whole
screen or parent window;  it means expanding the current window
so that the need to scroll horizontally or vertically is as
*    little as possible.


I hope the above comments are useful.  Although I reviewed the
material more than once, I may have missed or misunderstood
points that are already addressed.  In such cases, perhaps my
comments can reinforce a need for emphasis or clarification. 
Thank you for the open, inclusive, and systematic process that
has been evident in the development of these guidelines.

Regards,
Jamal

Received on Wednesday, 1 December 1999 23:10:10 UTC