Re: OBJECT and map

to follow up on what Jason White said:
> 
> A single, coherent solution is required, which integrates the need for a
> long description with the requirement for a textual alternative to the
> links provided by the image map. The OBJECT element provides the only such
> solution that is presently available or proposed, and I do not think that
> modifications to HTTP or image file formats of the kind that have been
> discussed in this group would serve as a satisfactory substitute. Thus, in
> answer to Daniel's question, I would suggest that loss of the capacity to
> refer to image maps by means of OBJECT would exclude a solution to the
> accessibility problem which, as argued above, not only satisfies the needs
> of users but is also likely to be more acceptable to authors than the
> redundancy introduced by LONGDESC and ALT text.
> 

I see some alternatives to USEMAP on OBJECT that we should be
prepared to discuss.

One is to embed a link to the long description of the image in a
client-side map of links.  [As with all MAP AREAs this would
require ALT text.]  This suggests that IMG+USEMAP can meet
minimum requirements, although OBJECT+USEMAP give better
capability to say things well.

The other approach is what, after my conversation with Murray at
the meeting, I would call metadata-related approaches.  This
involves getting accessibility needs integrated into a more
general approach to how information about resources is made
available.  It includes but is not limited to better use of
fields available in the HTTP response to a HEAD request.

Building on generic services for meta-information, and adding our
voice to the demand that these services work, ultimately involves
the lowest cost, in terms of extra effort by authors and
webmasters to achieve broad accessibility of their Web
utterances.  Note that a metadata approach _does not require_
changing the HTTP standard in any way.  But it requires
coordinated changes in server and browser behavior.  The USEMAP
on OBJECT strategy requires changes to browser and author
behavior.  Because the risks to success via metadata and via
USEMAP on OBJECT are different, there is a risk argument in favor
of pursuing both approaches.

For the non-image things that authors want to put in OBJECT
elements -- are we aware of accessibility problems?  How can
those be addressed?  Do these strategies allow us to meet our
need to link to a description of the image _inter alia_?
If we want to ask the HTML group to re-open the definition of
OBJECT, we should be ready with all problems with OBJECT that
we know about.

I think that HTML should implement the USEMAP on OBJECT, unless
there are good reasons against it that we don't know yet.  We
should communicate to the HTML group that we need a method of
providing links to long image descriptions that works for
sensitive maps, not just passive images.  That's why we liked
USEMAP on OBJECT but we are willing to listen to creative
alternatives.

I think the accessibility working group should simultaneously
pursue metadata-related approaches.  We should connect with the
people interested in the RDF topic and see how we can work
together.

--
Al Gilman

Received on Saturday, 16 August 1997 15:59:31 UTC