Re: MediaObject Proposal for LongDesc

["Dave Pawson" <dave.pawson@gmail.com>]

On 10/03/06, Richard Schwerdtfeger <schwer@us.ibm.com> wrote:
>
>
> Hi Dave, Chieko,
>
>  I read this and see value in it but it does not fully address an interoperability problem with Accessibility APIs.
>
>  In MSAA, Java, or ATK on Gnome there is the notion of an accessibleName an AccessibleDescription. Description would provide the equivalent of longdesc. It provides additional verbose text to the user when asked.


>
>  So in the case of an OK button:
>
>  The accessibleName is would be "OK" and is the equvalent of a label
>  The accessibleDescription might be "Activating this button will result in the indexing of your mail"

I agree on this as a general proposition.
However I am trying to improve on this for different modalities, by enabling
the provision of media other than text?

My basic content model for anything in a document not
in the same modality as the prime document (typically text)
is as below



>  <define name="med-mediaobject">
>    <element name="med:mediaObject">
>      <oneOrMore>
>        <ref name="med-imageobject"/>
>        <ref name="med-textobject"/>
>        <ref name="med-audioobject"/>
>        <ref name="med-videoobject"/>
>      </oneOrMore>
>    </element>
>  </define>

So rather than just insert an image/audio clip/video clip etc,
the user can insert  a number of alternatives, supporting
multiple modalities, which hopefully will match those of the reader.

I see this as an improvement of only allowing text as the
single alternative format for the richer media.

So, for Richs example, an audio clip, the author might well add
a text equivalent or a video clip which provide the same information.

It is my hope that this meets the needs that Rich describes.
I also believe it negates the need for either alt text or longdesc
but having read the usage, I'm less sure.
  A development might be to offer guidelines that the first block level
element in the textObject should be that 'brief' description, followed
by a more complete equivalent.



regards





--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk