Re: alt tags no with a twist

Colin Reynolds writes:

>IMO the only way to deal with this is to pressure the browser developers
>to remove the ALT-as-Tooltips functionality from their offerings. Check
>the www-html thread "Questionable implementation of IMG ALT attribute as
>tooltips" from the beginning of this year, in which Foteos Macrides
>astutely spotted that
>
>====
>   what you want is:
>        
>   <IMG SRC="foo.blah" ALT="alternative_string" TITLE="tooltip_string">
>
>   and broader implementation and understanding of this
>   provision in HTML 4.0.
>====
>
>For more info, check out the rant I wrote about this, at:
>
> <URL:http://www.the-net-effect.com/bad-tooltips.html>

	I agree that the use of ALT as tooltips regardless of the
TITLE tag hurts the web coming and going by 1) making sensible ALT
text look stupid as a tooltip and 2) causing people to define
tooltip-oriented ALT text which makes no sense in a text browser, like
ALT="BULLET".  But wouldn't a sensible strategy be:

	1) If TITLE="", no tooltip;
	2) If TITLE is anything else, use that as the tooltip
	3) If no TITLE, but ALT, use the ALT as a tooltip
	4) If neither TITLE nor ALT, use "[IMAGE]" as a tooltip (since
IMGs without ALT are wrong anyway)

That way, authors would provide ALT and then override it with TITLE if
it made a bad tooltip.  Sometimes a sensible ALT also functions fine
as a TITLE.  For instance, some images carry essentially non-textual
information, and the only reasonable ALT is something like ALT="[Photo
of the White House]".  Also setting TITLE="Photo of the White House"
seems redundant.

	Also, I'm not sure how one would get around the problem of
viewing ALT text for IMGs whose WIDTH and HEIGHT make them too small
for the ALT to fit, or for that matter for images that are already
loaded.  Perhaps that information could appear in a right-click menu.
(Since people who want to see the ALT text when the images are already
loaded are probably rare.)
					John T. Whelan
					whelan@iname.com
					http://www.slack.net/~whelan/

Received on Tuesday, 18 August 1998 17:29:39 UTC