WebAIM - Web Accessibility In Mind

E-mail List Archives

RE: WAVE 3.0 alpha

for

From: Jukka K. Korpela
Date: Jan 24, 2003 3:29AM

On Fri, 24 Jan 2003, Ben Coutts wrote:

> I read recently that best practice for spacer images was to set alt to ="*".
> The rationale was that the user knows that the image is meaningless rather
> than just having been forgotten about.

How would you like to listen to
asterisk asterisk asterisk asterisk asterisk asterisk asterisk asterisk
asterisk asterisk asterisk asterisk asterisk asterisk
when a page contains lots of small decorative images?

Why should a user know the difference between a meaningless image and
"just forgotten" image when he does not see any image?

There has been a lot of discussion about alt="" versus alt=" ". Instead of
going into all the details, and without actually knowing all the ways that
different browsers treat each of them, we can state that both of them
cause some problems on some browsers. The only sensible approach is that
authors write documents in a logical way, and user agents take it from
there. If there were some definite situation where the great majority of
browsers get one of the alternatives right and the other wrong, then we
could base the decision on that practice rather than logical
considerations. But there isn't.

The attribute alt="" specifies the empty string as the textual equivalent
to be used in place of the image. What could be more logical, if the
image is purely decorative? Using alt=" " could be sensible for _spacer_
images, since the space character might perform the same function as a
spacer image. But in fact most spacer images are completely irrelevant in
text-only presentation. If not, then it's often a _no-break_ space that
should be used. A few related notes:
http://www.cs.tut.fi/~jkorpela/html/alt.html#empty

And certainly even alt=" " is superior to alt="*" or alt="decorative
image" (to mention yet another approach, which has actually been
applied quite a lot!).

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


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