Re: ALT revisited

> > I've always thought ALT was under-specified in the HTML spec, which
> > only says something like: "For user agents that cannot display
> > images, this attribute specifies alternate text."
> > The "cannot" in the above sentence can be loaded with semantics..
> 
> I am struggling to bring the quality of the Cougar spec upto
> the expected level, and would welcome any suggestions for what
> we should include in the description of IMG, OBJECT and APPLET..
> 
> For IMG we have the ALT attribute, but could propose a further
> attribute for a link to further information about the image,
> e.g. ALTSRC. This has the effect of avoiding problems with down
> level browsers. We can readily extend Lynx and the emacs based
> browsers to support this new feature.

Adding an URL attribute to IMG is a way to handle the long description
issue. We're still waiting for the requirement report on that item to
see what's needed.

One of the requirement brought up at the meeting in Sophia was the
implementation-ness of the solution (can whatever we recommend be used
in today's browser). Extending HTML is of course not ideal, but having
rethought about other solutions, I'm not sure there is anything that
will comply with this requirement anyway (in particular, I realized
that the CSS 'display: none' feature is not in CSS1 core and therefore
is not available yet in all CSS-ready browser).


> For OBJECT, we can advise authors how to use the content of
> the OBJECT element to provide textual descriptions of images.
> Note that OBJECT supports use of the <A> element to define image
> maps. The anchor is extended with the SHAPE and COORDS attributes
> so that the textual description along with hypertext links also
> serves to define the hotzones on the image. For APPLET the course
> is less clear to me. The content of the APPLET allows you to
> include a textual description. There are also the guidelines for
> authoring Java in a way that minimizes accessibility problems.

> Any help would be much appreciated. We plan to publish the
> 1st draft of the Cougar spec in a few days time, so any input
> now could be folded into that draft. Otherwise, there will
> be time for further input before the spec reaches the proposed
> recommendation status.

Regarding this issue and others, we need to clearly differentiate what
belongs to the HTML specs (semantics of ALT vs. TITLE vs. ALTSRC
vs. content of A, etc) and what belongs in a browser UI guidelines
(e.g. how does the user activates the ALTSRC link of an IMG used as an
A).

Received on Monday, 30 June 1997 09:30:29 UTC