W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2000

Re: Links to Images

From: Gregory J. Rosmaita <unagi69@concentric.net>
Date: Thu, 04 May 2000 01:09:33 -0400
Message-Id: <>
To: Kynn Bartlett <kynn-edapta@idyllmtn.com>, "Gerald G. Weichbrodt" <gerald.g.weichbrodt@ived.gm.com>, <w3c-wai-ig@w3.org>
aloha, kynn!

Kynn Bartlett wrote,

Yes, a "raw image" linked from a page has accessibility problems; if you 
create even a minimal HTML page "wrapper" around it, you can add 
alternative text and other information.

with most of today's technology, kynn, you're right -- and, yet, 
technically, there's no reason why you shouldn't be able to point to a 
naked image if you utilize the LONGDESC attribute -- which, would, of 
course, necessitate support for LONGDESC by quote mainstream unquote 
browser manufacturers, and it wouldn't hurt if they were also Dublin Core 
[1] aware -- _then_ the user (no matter whom, no matter the type of device) 
could not only choose what he or she gets, but actually know (with 
confidence!) what one is using one's precious bandwidth to obtain...

support for LONGDESC would obviously obviate the need for a wrapper, if it 
is unnecessary or undesired, but would still leave the user with a 
multitude of options, such as: render the image and the LONGDESC as one 
document; display the content of the file referenced via LONGDESC in a 
separate viewport/window, and so on...

my conscience just tapped me on the shoulder to remind me that there are 
proxies to be built to deliver this sort of functionality on the fly to a 
request coming from older technology, such as an HTTP request from Lynx 
running on a shell account or someone running a commodore or atari that 
they bought on eBay, but i temporarily hushed it by reminding it that for 
embedded images, i use the "encase the image in a link that points at the 
URI referenced by the LONGDESC attribute" hack to expose the presence of 
long descriptions to the user by making the graphic itself a link to the 
URI referenced in the LONGDESC attribute set for the IMG element -- if, 
that is, the graphic warrants a long description -- sometimes ALT 
(especially when used in intelligent combination with the "title" 
attribute) is quite enough ...
speaking of LONGDESCs, a question i've been pondering is, should a document 
referenced using LONGDESC include a link to the quote naked unquote 
image?  and if not, why not?  isn't that providing an equivalent 
alternative to the descriptor?

of course, if the graphic was actually an XML based application -- marked 
up, in Scaleable Vector Graphics (SVG) [2] -- the image itself could not 
only be richly described, but distinct parts and components of that image 
could also be richly described, making it possible for users to get a lot, 
or a little, or none of the alternative information it is possible to embed 
in an SVG file -- depending, of course, upon the user's need or desire...

just one of the many reasons why an XML-based graphics syntax, such as SVG, 
is far superior to rasterized (binary) graphics, but i don't need to tell 
you that!


1. The Dublin Core: <http://purl.org/DC/index.htm>
2. Scaleable Vector Graphics: <http://www.w3.org/Graphics/SVG/Overview.htm8>
Received on Thursday, 4 May 2000 01:05:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:35:56 UTC