Re: launching the IG review

> DESC      Text associated with images

Here is my cut at the image description issue:

Images are very important to the look of most Web pages.
This naturally presents accessibility problems for people
with visual disabilities. Traditionally images in HTML pages
are specified using the IMG element, and to a lesser extent
with the form field INPUT element for image fields (type=image).

For both of these elements, authors can provide a short
description with the attribute named "alt". Most authoring
guidelines recommend authors to use this feature. HTML 4.0
makes this a required attribute.

Whats missing is an ability to specify a longer rich text
description for images -- a richer description that can
include headings, paragraphs, lists and hypertext links etc.
This is especially critical for images that are being used
for image maps.

HTML 4.0 introduces a new element named "object" for embedding
images and other objects into HTML pages. Unlike IMG, object
is a container that wraps around the alternative content to
be used when the object itself can't be rendered directly.
We are hoping that authors will switch from IMG to OBJECT over
a period of time.

Unfortunately measurements of which browsers people are using
show that many people take a long time to upgrade to newer
browsers. As a result, Websites are forced to design their
HTML pages to look good on older browsers as well as the latest
models.

As a result, the IMG element will be in use for a long time
to come, and we need to find a way for authors to make IMG
more accessible. The proposed solution is to add a new attribute
named "longdesc" that can be used to provide a URL pointing
to a longer description. This will in most cases be placed in
a separate file. People using say speech-based browsers will
first read the alt attribute and then choose whether to
follow the link provided by the longdesc attribute.

Some people have suggested other approaches. For instance,
including longer descriptions within the image data files.
Both GIF and JPEG have a provision for including text within
the image data. The downside is the long time it takes to
retrieve the image files, something most people will want to
avoid if they aren't going to be able to see the images anyway.

Another proposal has been to find some way to get Web servers
to look at the user agent info supplied with each request. When
a request for whitehouse.gif is received from a browser that
doesn't understand images, the server could be smart enough to
check its database to see if there is a description file for
this image. This sounds like a great idea! But experience has
shown that its hard to setup and maintain servers in this way,
both for Webmasters and for mere authors.

In the face of this, it seems like a reasonable idea to promote
the longdesc attribute for IMG. Its something that we can get
HTML editing tools to cater for and is easy to explain to authors.
People with regular vision win too, as often when the network
slows to a crawl, its convenient to turn off images, and soon
enough there will be a flood of small portable wireless devices
for browsing the web leading to increased emphasis on speech
output.

Regards,

-- Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
phone: +44 122 578 2984 (or 2521) +44 385 320 444 (gsm mobile)
World Wide Web Consortium (on assignment from HP Labs)

Received on Wednesday, 15 October 1997 12:53:31 UTC