Re: remembering ALTs that go with [SRC] URLs

AG2:: second round comments back from Al Gilman

At 10:54 PM 12/16/98 -0500, Harvey Bingham wrote:

>HB: I note that HTML and XML only allow one attribute of the same
>name in any one tag. Thus if some graphic had distinct multiple
>language descriptions built-in, it would be up to the application 
>to provide just the pertinent one for the language containing the
>reference to that document.
>
>HB: We have not proposed multi-language ATT values on a single
>graphic, that would require some delimited recognition and 
>selection of the desired single language message before its 
>presentation, say as a tool-tip. 

AG2:: Good point.  This came up briefly on WAI-PF talking about what
possibilities are opened up by XML.

One way to look at this is that ALT may be mis-cast as an attribute in the
first place.  I had a brief exchange with Martin Duerst on WAI-PF in which
the idea came up that it might make more sense to have the text which fills
the role now played by ALT be alternative _content_ and not subject to the
restrictions of _attributes_.  Then it would be easy to create alternative
structures with as much depth and breadth as we need, and choices can have
links and RUBY and phonetic equivalents etc. rolled into them.  Search in
the recent WAI-PF archives for "revolutionary."  In other words the
starting point for structuring things would be the smil:switch element and
not the html:alt attribute.

[...]
>><img-alt-use>
>><img-src>[URI]</img-src>
>><alt-text>[text]</alt-text>
>><as-used-in>[URI]</as-used-in>
>></img-alt-use>
>>
>HB: That approach will allow multiple <alt-text>[text]</alt-text>
>Is each part of a pair with its following <as-used-in>...?

AG2:: I was just thinking of a flat format of tuples.  Not that it has to
be that way.  This was just an example illustrating the use of named fields
so that later extension is possible without confusion.

Received on Thursday, 17 December 1998 15:53:32 UTC