Re: Desc and Alt

to follow up on what Daniel Dardailler said:

> Jason and myself have been looking at the issue Al calls "Text
> descriptions and alternatives for images" in his HC-plan page
> (
> We think we still need to add of something like:
>  In order to allow arbitrary anchor (such as long textual description)
>  in the OBJECT shapes content, OBJECT should only looks for <A> that
>  have a "coords" attributes and ignore the others when building the
>  image map.
> Even though we think this will be the case by default: no coord = no
> relevance for the image map and therefore no use.

I don't think that is browser behavior.  Tags that are not recognized
are ignored.  Content of unrecognized elements, on the other hand is
displayed because their containers are ignored and the content could
as well be text/plain.

> We think the MAP extension is not to be pursued, as it's not backward
> compatible for old browsers and would require the use of hacky comment
> techniques to gain support.

I don't understand why plain text as content in an OBJECT doesn't have
the same problem with old browsers as plain text in the content of
a MAP.  I hope Dave can clarify this point.

> We think we should ask the LONGDESC for IMG, the reasoning
> being that it's simpler to implement than OBJECT and so we should see
> implementation in Lynx/WebSpeak/Emacs sooner than expected and the
> benefit of the functionatily present for users.

I believe that for LONGDESC to be a solution, there must be some default
browser behavior defined.  I believe that there is not consensus on this
point, nor on what the default behavior is.   There is a smaller issue
as to whether the HTML spec or a separate document is the place to 
publish the default behavior.

> We do want FRAME to get a real TITLE attribute and a LONGDESC too, the
> reasonning being that FRAME becomes equivalent to IMG in that
> situation (result in inlining the image).
> Additionally, we think the note on generating ALT text automatically:
> should be improved to say that some prefix will be added to the
> generated alt text when the author is not responsible for it, so that
> the user know what quality to expect.
> We also think this section should be moved in the Browser guideline
> book, not in the HTML spec.

That point will generate more discussion, you will need to explain it.

-- Al

Received on Thursday, 9 October 1997 18:37:27 UTC