General comments on style sheets and WAI FP work

This is a convenient moment in which to introduce some general comments
regarding the future directions that can be taken in the WAI Formats and
Protocols effort.

If the WAI HC proposals are introduced into the HTML 4.0 specification
without significant amendments, then I think the success criteria for the
HTML review will have been met.

CSS 2 will be of significant benefit in the development of a more
accessible web, since its widespread adoption will discourage authors from
abusing the HTML markup in ways that are detrimental to the creation of
braille and audio representations of a document. Furthermore, the audio
style sheet mechanism included in CSS 2 will significantly improve the
quality of audio rendering, provided that it is widely implemented by the
developers of assistive technology and audio-based user agents.

My suggestions regarding the audio formatting of tables remain to be
considered in detail. A means of controlling the rendering of ALT text (by
associating audio or visual style properties with it) is also needed.
Furthermore, the CSS 2 review should examine the question of whether
further style properties are needed to enable user agents better to
indicate the presence of a long description associated with a graphical
resource (see the LONGDESC attribute proposed by the WAI for inclusion in
HTML 4.0).

Media types for braille and TTY terminal devices have been proposed to the
HTML working group, but the associated style properties have not been
defined. I agree with Al that braille devices and TTY terminals may be
able to share some property definitions, but the analogy should not be
taken too far. Item numbering, footnotes and end notes, and related
concerns identified by the WAI have not as yet been addressed in CSS.

More generally, I think there is an important strategic decision which the
WAI Formats and Protocols Working Group will soon be obliged to make,
namely, how far to push the development of CSS, or whether the needs of
accessibility would better be served by concentrating principally on other
style mechanisms, such as the XML style language, which will reputedly be
based on DSSSL. The XML style language is not even at the draft stage, a
fact which can be both an advantage, in that it may be possible for the
WAI to influence its development at a very early stage, and a
disadvantage, in that a final specification and a working implementation
thereof is unlikely to be available in the near future.

The ability to transform the structure of an HTML document by reorganising
its elements is an important aspect of any style mechanism that can
adequately meet the needs of braille and audio users. So far, there has
been no indication of an intention to include such tree transformation
properties in CSS, and as T.V. Raman has suggested, this may be a good
reason for concentrating the WAI effort more on DSSSL and the forthcoming
XML style language. More information is required from the CSS and XML
working groups in order that the WAI can make an appropriate decision in
this regard.

XML accessibility is a problem which has so far not been given detailed
treatment. Following the model of the ICADD mechanism, I would suggest
that a means should be developed whereby the developers of XML DTD's can
map them onto the HTML 4.0 DTD, or a subset thereof. Given that HTML will
be widely implemented and that, if WAI enhancements are accepted, it will
provide excellent accommodation for the needs of people with disabilities,
it would be preferable to map XML elements to corresponding HTML markup
than to rely on the now obsolete ICADD 22 DTD. I understand that George
Kerscher is already working on this problem.

As mentioned above, much effort needs to be devoted to the development of
a braille style sheet system which is adequate for the conversion of HTML
and XML documents into well formatted braille. Again, the direction which
this work should take depends on the over-all approach to CSS and the XML
style language which the Formats and Protocols Working Group adopts.

HTML mathematics is another very important specification that warrants a
detailed review, in order to ensure that braille and audio representations
can be automatically generated from a well marked up mathematical

The advantages of RDF as a linking mechanism need to be explored,
especially in respect of phonetic and abbreviation dictionaries,
descriptions of tables and images, etc.

Finally, none of the foregoing projects is likely to achieve its full
potential in practice unless an adequate interface, based for instance on
the Document Object Model, between user agents and assistive technologies
(such as braille and audio output software) is developed and implemented.

Thus, it is vital that the Formats and Protocols working group develop a
plan for addressing each of these issues, and that a sufficient number of
experts, with relevant knowledge and insight, be brought together to carry
out the various projects which such an effort necessitates.

Received on Saturday, 8 November 1997 19:34:58 UTC