Bridging the Discourse Structure/Presentation Structure gap

Somehow we need to take the "magic" step from the (media independent)
discourse layer to the (media dependent, but device independent???)
presentation structure layer.

What do we mean by Presentation Structure?

Grouping of media items (and groups of media items)
Ordering (in a space/time/link neutral sense)
But there is also something about media type - 


Communicative device layer

Input:
CommDevice tree
Content = number of media items,
the media items themselves (implicit width/height/duration),
Rhetorical structure guiding the choice of the device

Context:
Where the communicative device fits in the hierarchy of communicative devices.

The "rules" of the device:
Need to know attributes of media?  Can we restrict it to attention
demands (e.g. can't mix video with text)?


Output:
Presentation structure



[[Note for Lynda.
In AHM you can hang temporal and atemporal composites all over the
place.  There are no restrictions - the composition is only dealing
with the temporal dimension. Channels deal with spatial hierarchical
composition. With CommDev it is unclear what you can hang where.]]

There are N-S and N-N CommDev's.


Can we actually keep it simple?
Take the RST hierarchy.  Show as many levels as possible at once.
Then only later worry about video/audio/image restrictions?
Add them in as constraints?



Video is not stretchy.
Can't show two videos at once
(Unless closed loops, see McCloud).


Images can be stretchy, but not often in aspect ratio.

Text can be reshaped.

Audio shouldn't be played too many at once.
   Background music
   text commentary
   Sound effects are more than enough


Professionals can break the rules, but we can't.


Can't have rules about whether video goes before text.

_Can_ have rules that important information goes into text or images,
since video is transitory.  Although processes are difficult to see in
images.

Need to know whether we have a description of a process.