Re: Thoughts on WCAG 2.0 {3.2}

> It actually has as much to do with the structural aspects that are
> implicitly generated by classes [...] This is almost never the case
> because scanning visually for red stuff is a technique that clearly
> provides an advantage for the retinally-enabled (blindlesss) reader.

True, and therefore, the fact that a certain portion of XML/HTML was written
by Robin is important both semantically and presentationally. The whole
point of {3.2} is that style should be added to enhance both the structure
and the semantics.
For example, say you had the following fragment:

     <p dc:author="Robin">This bit was written by Robin</p>
     <hr />
     <p dc:author="Sean">This bit was written by Sean</p>

What does the <hr /> stand for if a blind user was looking at it? It is in
fact a border between one author of the document and another. But having <hr
/> doesn't imply that, it's just a visual horizontal rule: and *that* is one
of the greatest problems with HTML. You should write:

     <p dc:author="Robin">This bit was written by Robin</p>
     <hr title="Division between content implying different authors" />
     <p dc:author="Sean">This bit was written by Sean</p>

But that still doesn't mean much :-)
In reality, because <hr /> is presentational and only for screen media, you
should scrap <hr /> altogether, i.e.

     <p dc:author="Robin">This bit was written by Robin</p>
     <p dc:author="Sean">This bit was written by Sean</p>

And use the following CSS style sheet:

@media all {
*[dc\:author="Robin"]:before { content: "The following section was written
by Robin: "; }
*[dc\:author="Sean"]:before { content: "The following section was written by
Sean: "; }
}
@media screen {
p[dc\:author="Robin"] { border-bottom: 1px solid #000000; }
p[dc\:author="Sean"] { border-bottom: 1px solid #000000; }
/* Really this should apply to all block level elements,
but I just used p as an example */
}

But how many peole would do that? None, becuase most don't know how to, and
Dublin Core isn't even allowed in HTML anyway yet...we would have to use
title="Author:Robin" or class="byrobin", which is just a hack: and we must
recognize it is *only* a hack, not the proper method. The proper method is
to do what I have done above! (Where the semantic meaning is given first in
the @media all section, and then the presentation is tacked on in the @media
screen section.)
The point I am making is HTML is a prenentation based language, with very
little room for accessibility features. However, if you are *very* careful,
you should in the very near future be able to mark up your document
semantically:

     1. DON'T use class="[...]" to imply semantics!
     2. Instead, use proper metadata languages, and combine that with CSS

> [...] the act of creating/using a class in CSS involves imparting
> content/structure (in addition to presentation) and it has, by
> nature of the way CSS is used/revealed, enabled abuse in
> accessibility terms.

I'm glad you follow...class="[...]" is a bad idea, becuase it doesn't say
anything about that class most of the time. Things like class="emphasis" and
so on are O.K., but even then you sould use proper meta/HTML markup.
In summary, I believe Checkpoint {3.2} in WCAG 2.0 is a good one, but it
requires a lot of explanation, and techniques/implementation to make it
work. There should also be a mention that misuse of structural markup to
imply semantic meaning is *highly* inadvisable, and that styling based on
semantic markup is encouraged.

Kindest Regards,
Sean B. Palmer
http://xhtml.waptechinfo.com/swr/
http://www.w3.org/WAI/ER/
"Perhaps, but let's not get bogged down in semantics."
   - Homer J. Simpson, BABF07.

Received on Thursday, 16 November 2000 10:08:39 UTC