W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2000

Re: screen readers, browsers, & the reporting of ALT on & in image maps

From: Alan J. Flavell <flavell@a5.ph.gla.ac.uk>
Date: Wed, 12 Jan 2000 12:32:09 +0000 (GMT)
To: "Gregory J. Rosmaita" <unagi69@concentric.net>
cc: WAI Interest Group Emailing List <w3c-wai-ig@w3.org>, WAI Guidelines List <w3c-wai-gl@w3.org>
Message-ID: <Pine.OSF.4.20.0001121208430.23730-100000@a5.ph.gla.ac.uk>
On Wed, 12 Jan 2000, Gregory J. Rosmaita wrote:

> 4. Lynx32 renders the ALT text Picture of a Compass as a hyperlink (which
> suggests that a better textual equivalent could be "Picture of a Compass, used
> as in image map" or (my preference),

This is a delicate point, but I'd suggest as a prima facie assumption
in web design (this was expressed early on in web style authoring
advice at the W3C, or might even pre-date it) "don't mention the
mechanics".  Sometimes, it's true, one has to deviate from this
"ideal", but doing so often causes more difficulties in other browsing
situations than it solves in the situation(s) that the author had in

_If_ this imagemap is going to be entirely serviceable as a list of
links (which is how it's presented in Lynx), then why would it be
necessary to pester the non-visual reader with the fact that the
original was designed as an imagemap?  And, in the case which you are
presenting, I'd suggest that it is indeed entirely serviceable in that
way, at least when used with browsers that are capable of that kind of
usage.  If, on the other hand, it cannot be practically used in that
way (for example, where the original is a street-plan whose only
purpose is to exhibit geographical relationships) then of course some
other approach is needed.

> "Select A Direction"

Now, that seems an excellent solution, to me.  As an ALT text, it
represents a functional _alternative_ to the original visually-based
design. Any reference to the visual appearance of the non-viewed
image, if required, would be better delegated to the TITLE and/or
LONGDESC attributes (traditionally: D-link).

That seems to me to be the correct _authoring_ approach. Browsers then
stand to be rated on their ability to support this alternative
function when images cannot be viewed.

(As an aside - there's no reason why every web browser on the market
has to do a good job of handling every kind of usage.  There's a niche
for specialised browsers that concentrate on supporting particular
pattern(s) of usage.  For example, you wouldn't expect WebTV to be an
effective speaking browser, any more than you'd expect a text browser
to display images on a character-cell display.)

> when one does follow the image map link, the four AREAs are listed in an
> ordered list
>         * [1] North
>         * [2] South
>         * [3] East
>         * [4] West
>  on a page whose TITLE is the ALT test defined for the image "Picture of a
> Compass",

"Picture of a Compass" might well, as I remarked above, qualify as a
TITLE attribute for this IMG element, but as an ALT (i.e _alternative_
text to be used when the picture is not being presented visually) it
seems to me to be counterproductive, as it distracts attention towards
something which the reader not only cannot see, but cannot use in that
way - and which they do not _need_ in order to be able to use the
facility in the _alternative_ text-based manner.

best regards
Received on Wednesday, 12 January 2000 07:32:12 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 16 February 2017 16:48:24 UTC