RE: longdesc media type concern

For my part, the issue is not a question of how the LONGDESC attribute
should be used, but of where the constraints should be specified. The
argument which Al and others have been running is to the effect that
the WAI guidelines, (in particular the Techniques Document of the Web
Content Accessibility Guidelines) offer the appropriate context in
which to constrain the usage of LONGDESC. Accordingly, it has been
argued that the internet media (MIME) type of the resource to which
the LONGDESC URL refers, should not be formally restricted as part of
the HTML and XHTML language specifications. For example, do we have a
guarantee in advance that all document markup languages, based on XML
for example, will be subtypes of "text" (E.G. "text/xml",
"text/xhtml", etc.)? If not, then there are good grounds on which to
conclude that the best means of regulating LONGDESC does not consist
in specifying a constraint in terms of the internet MIME type of the
destination resource. The central claim here, is that the hasty
imposition of a limitation based on MIME types may yield adverse
consequences in the future, as it depends on certain assumptions
concerning (1) the range of acceptable uses of LONGDESC, and (2) the
future of the MIME type system, which have not been fully explored.

As an alternative, the language specifications could make it clear
that text, and more specifically a marked up document providing a
description of the image, should lie at the destination of the
LONGDESC URL. Moreover, readers should be referred to the Web Content
Accessibility Guidelines (or the Techniques Document, as this is where
the issue will be discussed), for details pertaining to proper usage,
etc.

Since this language would be informative only, it would seem feasible
to refer to a W3C note such as the Techniques Document, even though
the latter is not a normative reference.

Received on Monday, 6 September 1999 01:09:01 UTC