RE: @longdesc - starting over (was RE: Acessibility of <audio> and <video>)

Lachlan Hunt wrote:
> 
> Yes, this is a chicken and egg problem.  But it's not a problem the
> spec can solve. 

Yet it's a problem the spec seeks to ignore by removing the attribute?

> It's a problem the accessibility community needs to
> solve, and then come back with the evidence if and when it is solved.
> Until that time, longdesc hasn't got a chance of being included in
> HTML5, whether you like it or not.

It is gradually being solved now, as you note below:

> Now, I have been informed that that beta of JAWS 10 turns on the
> announcement of longdesc by default. 

More evidence that progress is being made:

	"The alt attribute is no longer displayed as the image tooltip when
the browser is running in IE8 mode. The target of the longDesc attribute is
now used as the tooltip if present; otherwise, the title is displayed. The
alt attribute is still used as the Microsoft Active Accessibility (MSAA)
name. If this attribute is not present, the title attribute is used
instead."
[http://msdn.microsoft.com/en-us/library/ms534132(VS.85).aspx]
(** the significance of this I do not think has been fully appreciated! **)

	"iCab [provides access to @longdesc] directly through the contextual
menu's Image->Description
(which opens the @longdesc URL in a new window). In Gecko-based browsers you
need to, through the contextual menu, open the image's "Properties" window,
which lists the @longdesc URL as a "Description". In the current Firefox the
URL is just text, but I'm told that in Netscape 6 and 7 it
is a (clickable) hyperlink."
[http://tinyurl.com/6ryq4u]
(I have neither installed on any of my machines - can someone else confirm
or deny this statement? - JF)

	"In the last two months I've come across more examples of the
LONGDESC attribute in use, than I've seen in as many previous years."(Date:
Tuesday, 12 Feb 2008 18:13)
[http://tinyurl.com/25klet]
(In the interest of transparency, this article does question the whole
subject of @longdesc, and echo's back information that we already know via
Hixie's stats and other conceded "issues" - what is interesting and germane
however is that the author has noticed a remarkable upswing in the usage of
@longdesc)

	"I work for a financial information company and a few of our older
pages (from before I started) have D links. Horrible things. We now use
longdescs on our charts pointing to the dynamically generated versions
containing accessible tabular data." - Robin 
[http://tinyurl.com/9hayv]

- lack of native support for @longdesc in WebKit is considered and tracked
as a bug: 
https://bugs.webkit.org/show_bug.cgi?id=10448

- lack of support in Opera has been called a bug as well:
http://www.gtalbot.org/BrowserBugsSection/Opera9Bugs/

- evidence of individual developers trying to fix or improve support of
@longdesc (band-aids):
  * Longdesc Linker for Internet Explorer 6:
http://www.hackcraft.net/longdesclink/
  * Firefox Longdesc Extension:
https://addons.mozilla.org/en-US/firefox/addon/273
  * Longdesc Acid Test: http://www.malform.no/acidlongdesctest/

************

One thought that occurs to me is this:  Since 

A) all of the major browsers are now moving towards better implementation of
@longdesc, and
B) the mainstream AT tools are getting better at supporting @longdesc, and 
C) for backwards compatibility reasons the mainstream browsers will most
likely not go "backwards" in their support of @longdesc, 

...why would the next generation spec simply ignore this attribute? Since
some authors will continue to use an attribute that *has* support in
user-agents and AT, you are now setting up a scenario where using a useful
(supported) feature will render a page non-conformant due to the use of a
deprecated attribute that currently has no alternative to this
functionality.  Say what?  Would it not seem more reasonable to provide
support and better guidance of usage, which is one of the more positive
aspects of the HTML5 draft?

Finally, in my previous post I posed a legitimate question: if not
@longdesc, then what?  This question remains unanswered to date. Until that
time, longdesc is the best choice to be included in HTML5, whether you like
it or not. 

JF

Received on Saturday, 6 September 2008 21:11:19 UTC