Natural Associations [was Re: Illustration]

[A PF matter, with great relevance to GL issues.]

In the XML Accessibility Guidelines (XML GL) currently, we have the
following guideline and checkpoint:-

[[[
Guideline 1 Ensure that authors can associate multiple media objects
as alternatives
1.1 Make sure this is done in the most natural way possible [ignore
the @@ and the priority].
]]] - http://www.w3.org/WAI/PF/XML/gl-20010422

At the moment, PF is undertaking to edit this guideline from its
current obfuscated form into something with a little more impact. In
doing so, we have found that it overlaps with some issues that GL have
necessarily appropriated to themselves.

In our case, we have the guideline "Ensure that authors can associate
multiple media objects as alternatives" ... what we are saying is that
as far as the schemata for data-oriented applications (see the
document for a definition) are concerned, we want to enforce the
ability to tie in some repurposable equivalent alternative with some
content who's semantics may not be repurposable.

Typically, this is represented by the age old "alt" text example in
XHTML, viz.:-

   <img src="banner.gif" alt="Welcome To My Site" />

However, this can clearly no longer be taken as an acceptable example
by itself, because it's a subsumption of the complexity of
relationships that occur between content of different modalities.
Thus, what I (we) are asking is for rationalization of the above where
appropriate (e.g. in guidelines) - the fact that it's a subsumption
should (at least) always be noted, and preferably) be expanded upon.

For an example of what we want to achieve, let's take the XHTML
"problem" again. What the alt text example says is that textual
alternatives must be provided as an annotation for the image... in
other words, we are thinking in terms of "I have an image, and now I
have to provide some equivalent alternative in the form of an alt
attribute". It is quite possible, and in fact often required for some
content authors to annotate a concept, a run of text, or marked up
content with an image. This isn't possible in XHTML 1.0/1.1/m12n
(which is a great shame), but should be in 2.0, so for example:-

     The starving babies were in a desperate state.

could become:-

     <illust src="babies.gif">The starving babies were in a desperate
       state</illust>

Note that this is not simply limited to illustrations - the ability to
associate a variety of media alternatives no matter what the modality
should be pervasive throughout all XML data-oriented applications.

The question thereafter becomes about UI: how do we associate the run
to the image, where do we put that image, what do we do with the
image? That's beyond the scope of XML GL (although may be in the scope
of GL), but it could be that using some remote styling mechanism, one
can explicitly style where they want the content to appear, and
exactly how it should be styled. cf. 2.1 in XML GL. There's also the
question of multiple illustrations for one run of text... note how
much more interesting it becomes when (for example) you add the images
to the text rather than the text to the images.

So that's one small aspect of what we mean by natural. Another is the
question of what we mean by repurposable images themselves, for
example SVG. One view of the relationship here is that SVG enables
people to export bits of semantics from the image itself. Others may
prefer to think of it as the old annotation fashion again, i.e. adding
the descriptions to the image as a secondary effect. This is not
something that languages constrain upon people - it allows us to
choose, which is correct.

Hence, a "technique" for what we mean by "natural association" is a
mixture of both not being constrained by current technologies as to
the choice of semantic association mechanisms (and hence removing
restrictions on the form of the language itself, and removing
accessibility barriers).

As for checkpoint 1.1 (soon to be 1.3 due to document reorganization)
itself, I propose that WAI PF change the text to the following:-

1.3 Provide unconstrained mechanisms for semantically associating
content of different modalities with one another. That is to say, if
the semantics of part of a document instance are unlikely to be
accessible to all of your readers, provide equivalent alternatives
which facilitate understanding by other means.

(Note that I'm using a modified earlier proposal for the explanatory
text from Dave Pawson after the "-", which I found excellent).

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .

Received on Tuesday, 15 May 2001 13:55:28 UTC