WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: ALT tag vs. LONGDESC

for

Number of posts in this thread: 4 (In chronological order)

From: Kevin Price
Date: Thu, Mar 07 2002 1:12PM
Subject: ALT tag vs. LONGDESC
No previous message | Next message →

The W3C guidelines don't make it clear to me when exactly to use an ALT tag
only or when to use the LONGDESC tag. It seems like there should be a
number of character limit for an ALT tag. Is there any rules that people
are using to determine whether LONGDESC should be used? It appears to me
that even when you use LONGDESC, an ALT tag should also be associated to
that graphic. (Especially because only the very latest browsers support
the LONGDESC tag)
thanks,

Kevin Price
Coordinator-Assistive Computing Resource Center (ACRC)
Academic Computing and Media
California State University, San Bernardino
5500 University Parkway / PL-002
San Bernardino, CA 92407-2397
909-880-5079 FAX 909-880-7075



---
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/


----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/

From: Terry Brainerd Chadwick
Date: Thu, Mar 07 2002 1:36PM
Subject: Re: ALT tag vs. LONGDESC
← Previous message | Next message →

Kevin wrote:
>The W3C guidelines don't make it clear to me when exactly to use an ALT
>tag only or when to use the LONGDESC tag. It seems like there should be
>a number of character limit for an ALT tag. Is there any rules that
>people are using to determine whether LONGDESC should be used? It appears
>to me that even when you use LONGDESC, an ALT tag should also be
>associated to that graphic. (Especially because only the very latest
>browsers support the LONGDESC tag)
>thanks,

Bobby, in its Section 508 version, has been warning people to not use the
ALT attribute for anything over 150 characters.
"ALT text >150 characters, consider providing a separate description."

This is just a guideline however. On the sites where I've been given this
warning, the text in the alt attribute have conveyed the exact text that
was in the image so I didn't see any reason for a longer description. (I
did recommend that my clients drop the image in favor of real text.)

Terry

Terry Brainerd Chadwick, InfoQuest! Information Services
mailto: = EMAIL ADDRESS REMOVED = 1-503-228-4023 http://www.tbchad.com
Providing Internet Audits and Optimization to Improve Performance
Accessibility, Analytics, Content, Promotion, Search, Usability


----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/

From: John Foliot - bytown internet
Date: Thu, Mar 07 2002 3:01PM
Subject: Not versus, but plus (was: ALT tag vs. LONGDESC)
← Previous message | Next message →

Kevin,

EVERY <img> requires the ALT attribute, some *May* also require the LONGDESC
attribute on top of the ALT attribute. If the image is conveying
information beyond what would be on the surface, then employ the LONGDESC
attribute which links to a text file with a fuller explanation.

For example, you have a web page with an image of a pie chart:

<img src="piechart.gif" height="300" width="300" alt="pie chart showing
browser usage statistics" longdesc="/longdesc/piechart.html">

For non-graphic browsers, the alt tag is describing "what" the image is,
however, should the user wish to "see" what the percentages are, they would
follow the LONGDESC link to "piechart.html" which would be a text version of
the browser usage stats (it could be a simple .txt file as well).

To my knowledge, currently only IBM's HomePage reader supports the LONGDESC
attribute, so you may also see the "d" link used. While not a W3C
recommendation (nor I believe Section 508), some advocacy groups propose it
as a stop-gap measure while waiting for the browsers/assistive software to
catch up:

<img src="piechart.gif" height="300" width="300" alt="pie chart showing
browser usage" longdesc="/longdesc/piechart.html"><a
href="/longdesc/piechart.html" title="text description of user stats">d</a>

Some feel that the d-link is "ugly", but you could set the colour of that
link as the same colour as your page background (when browsers universally
support the visibility attribute of CSS2 you could use that as well):

<a href="pichart.html" title="text description of user stats"
style="visibility: hidden; color: White; font-size: 1pt;">d</a>

<opinion>
While normally I would say that we should never set font sizes as absolutes,
I'm also a believer of exceptions to every rule, and this would be one of
them. It could be argued that this may remove the link from text only
browsers, but generally they rarely (if ever) support style sheets anyway,
and we all need to accept the fact that aesthetics DO play a role in our
work. The d link is not mandated anywhere, it's inclusion is simply one
more assistive coding device that web developers may apply, and speech
browsers/screen readers will still "hear" this link. It's a balance thing,
and IMHO I've added more than I've removed...
</opinion>

The bottom line however is the use of LONGDESC is a judgement call, which is
one of the reasons why an automated "compliancy tester" (i.e.: Bobby or
equiv) is never the final word... it's a great testing tool, but not the
only one you should use (I believe even Bobby's web site says something to
the equivalent)

HTH

JF






>

From: Kevin Price
Date: Thu, Mar 07 2002 3:33PM
Subject: Re: Not versus, but plus (was: ALT tag vs. LONGDESC)
← Previous message | No next message

John,
Thanks for your comments.
JAWS 4.01 reads LONGDESC tags now in Internet Explorer 5.x and 6.
Kevin Price

At 05:01 PM 3/7/2002 -0500, you wrote:
>Kevin,
>
>EVERY <img> requires the ALT attribute, some *May* also require the LONGDESC
>attribute on top of the ALT attribute. If the image is conveying
>information beyond what would be on the surface, then employ the LONGDESC
>attribute which links to a text file with a fuller explanation.
>
>For example, you have a web page with an image of a pie chart:
>
> <img src="piechart.gif" height="300" width="300" alt="pie chart
> showing
>browser usage statistics" longdesc="/longdesc/piechart.html">
>
>For non-graphic browsers, the alt tag is describing "what" the image is,
>however, should the user wish to "see" what the percentages are, they would
>follow the LONGDESC link to "piechart.html" which would be a text version of
>the browser usage stats (it could be a simple .txt file as well).
>
>To my knowledge, currently only IBM's HomePage reader supports the LONGDESC
>attribute, so you may also see the "d" link used. While not a W3C
>recommendation (nor I believe Section 508), some advocacy groups propose it
>as a stop-gap measure while waiting for the browsers/assistive software to
>catch up:
>
> <img src="piechart.gif" height="300" width="300" alt="pie chart
> showing
>browser usage" longdesc="/longdesc/piechart.html"><a
>href="/longdesc/piechart.html" title="text description of user stats">d</a>
>
>Some feel that the d-link is "ugly", but you could set the colour of that
>link as the same colour as your page background (when browsers universally
>support the visibility attribute of CSS2 you could use that as well):
>
> <a href="pichart.html" title="text description of user stats"
>style="visibility: hidden; color: White; font-size: 1pt;">d</a>
>
><opinion>
>While normally I would say that we should never set font sizes as absolutes,
>I'm also a believer of exceptions to every rule, and this would be one of
>them. It could be argued that this may remove the link from text only
>browsers, but generally they rarely (if ever) support style sheets anyway,
>and we all need to accept the fact that aesthetics DO play a role in our
>work. The d link is not mandated anywhere, it's inclusion is simply one
>more assistive coding device that web developers may apply, and speech
>browsers/screen readers will still "hear" this link. It's a balance thing,
>and IMHO I've added more than I've removed...
></opinion>
>
>The bottom line however is the use of LONGDESC is a judgement call, which is
>one of the reasons why an automated "compliancy tester" (i.e.: Bobby or
>equiv) is never the final word... it's a great testing tool, but not the
>only one you should use (I believe even Bobby's web site says something to
>the equivalent)
>
>HTH
>
>JF
>
>
>
>
>
>
> >