tech-specific checkpoints

"CSS
Guideline 1.1 Provide a textual equivalent for every non-text (auditory
or graphical) component or multimedia presentation.

For CSS, this applies to use of color? Generated text? Aural cascading
style sheets? Does it apply at all?"

WL: I think this is where we must deal with the point raised about those
of us who tend to use presentation semantically through the side door of
making a "sounds meaningful" class distinction among stylized elements.
The old .warning class of <EM> situation.

"HTML
Guideline 1.1...
Provide any of this information within the content of the document."

WL: Of course this latter technique is applicable to the CSS one above.
Unfortunately most authors will feel put upon to be urged to use this
for *everything*. First they're given a method of styling that is
supposed to help with the problem only to find that they still must
explain that this particular style variety is being used to convey some
detail of the semantic intent. "You mean that if I want to have degrees
of emphasis that I must explain them instead of just showing them?" 

"Mark up lists and list items properly" is probably also needed in the
CSS examples because there is always the temptation to use <LI>, etc. as
CSS entries in order to convey gradations of semantics. 

I suppose we could write an essay on just how dependent we blindless
people have become on using presentation expressively. Although we like
to pretend that there's "pure" presentation items, in fact *most* uses
of "style" are reflections of voice inflections, etc. that aren't a
regular part of printed text. The question mark and other punctuations
are actually markup elements that direct one to read the passage aloud
in a certain way - flying in the face of the notion that presentation
and content aren't joined at the hip.

"Avoid using images to represent text -- use markup and style sheets
instead" pulls my chain almost as much as it will J. Chetwynd's! If what
is meant is to "Avoid using images OF text" I understand it but I think
"representing" is more like "illuminating" than "replacing"? In fact in
some cases we should probably be encouraging image use in connection
with text representation - after all text is actually a visual
represntation of speech.

"Guideline 2.3 Where presentation is used to communicate distinctions of
meaning or structure within the content, ensure that the meaning is
captured in the markup."

WL: As noted above the question becomes *HOW?*. The markup isn't
sufficiently "rich" to capture the nuances of meaning in many cases - it
would be too cumbersome, I think. When you read any of these sentences
aloud note how their semantics is affected merely by which word you
emphasize. If I were to emphasize "you" in the previous sentence it
would be interpreted to mean as a statement about how this is a function
of a particular person whereas if I emphasize "semantics" I imply that
other features than their meaning are not affected, etc. etc. This would
be tedious to do in markup.

Guideline 2.5 Ensure that distinctions needed to the render the content
auditorally or tactiely are captured in the markup."[there's a
gratuitous "the" in there]

WL: Again more guidance on how the markup can capture these
distinctions. The technique of elucidating these nuances in content
might be hard to sell.

"Guideline 3.4 Divide large blocks of information into groups where
natural and appropriate."

WL: Each of these guidelines might require an extensive tutorial in
what's "natural and appropriate"? Very few authors understand the
reasons for structure: "I'm not going to have a ToC!"

"Guideline 5.1 Make sure that the use of newer technologies transform
gracefully when a browser does not support the technology or the
technology is not enabled in the user's configuration."

WL: I've always been troubled by "transform gracefully" which has a nice
ring to it but if I'm not mistaken what it means is that whatever you're
trying to do disappears without messing things up when the browser
provides no support for it? 

ENOUGH!

Received on Wednesday, 23 August 2000 12:52:38 UTC