WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Lists of links to alternative formats

for

From: Jukka K. Korpela
Date: Aug 19, 2006 5:20AM


On Fri, 18 Aug 2006, Paul R. Bohman wrote:

> ... But the part I'm most interested in is how to represent links to
> alternative formats.

That's a tricky problem, but let's consider the users' needs first. If
there is a collection of documents in different formats, the user may wish
to act in two ways:
1) select the document first, then the format
2) select the format(s) first, if some formats are either strongly
preferred or strongly avoided.

The former is more common. The latter could be handled by providing some
tools, like alternative lists or some dynamic (scripted) tool for
removing some formats or selecting a particular format.

Selecting the document first is easier if there is no other selection that
_could_ be made on the page. People with cognitive disabilities will
probably find it easier to make one choice from a list than a
"two-dimensional" choice.

Thus, logically, you need a simple list of documents, or perhaps nested
lists or a table, but anyway organized by document content only, and
naturally using descriptive document names. In a simple approach,
selecting a document would then open a new small page containing links to
the different versions of that document. Naming the links would not be an
issue any more; you could use even format names, or perhaps a little more
logically "HTML version of Introduction to Foo Bar", "PDF version of
Introduction to Foo Bar", etc.

If opening a new page seems too clumsy, you could set up a piece of
JavaScript that opens a menu within the current page, with a link to the
small page as backup for non-JavaScript browsing situations.

This approach would simulate the ideal (?) solution where document format
selection takes place automatically by user preferences. The ideal
solution is technically quite possible as far as the protocols are
considered: you could set up content negotation in HTTP, so that a browser
would request for a document using a URL like
http://www.example.com/foobar and send a suitable Accept header,
specifying the acceptability of different media types, and the server
would then pick up the media type (among the versions it's got) that best
fits the request. However, this fails for two main reasons: browsers send
a more or less fixed and often nonsensical Accept header (e.g., IE says
that it will accept any media type, with no expressed preference between
them), and even if the allowed users to customize such behavior, most
users would not do so (as we can see from the fact that Accept-Language
headers, which _are_ customizable by users, have very rarely been actually
selected or even checker by users).

But this can easily be simulated using a link to a list of alternate
formats (which is more or less what content negotiation is expected to
do as a fallback, i.e. when none of the available formats is acceptable
according to what the browser says).

--
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/