> About FIFO versus LIFO, I believe you are correct. I did several of these
> things FIFO but I never did use them much.

Actually, none of them is correct. Take this Josquin example,
for instance (imagine treble clef):



    ---------------------------------------------+----------
	  +==============+                       |
    ------|----|----|----|----------------+==+---+----------
	  *    |    |    |    |           +==+   |
    -----------*----|----|----|-----------|--|---+----|-----
	  o.        *    |    |           |  |   |    |
    ------|--------------*----|-----------|--|---+--o-|-----
	  |                   *.      *   |  |   |  | |
    ------|---------------------------|---*------+--|-o-----
	  |                           |      *      |
				      |             |

		   (A)               (B)


the a4 at (A) needs to be turned off in a stack-like (LIFO) manner, 
whereas the f4 at (B) requires a FIFO interpretation of incoming
note-offs.

This dilemma cannot be solved without looking at other performance
parameters, such as attack velocities, release velocities, average
overlap of note-ons and note-offs in a run to determine the degree
of legato, etc.  Very messy.

Of course, since most controllers don't allow you to play a new 
attack before the note is released (like grand pianos do), this 
is only an issue, if two keyboards are used on the same channel.

In message <379E40BD.5C388E0C@ccrma.stanford.edu>you write:
>
>> About FIFO versus LIFO, I believe you are correct. I did several of these
>> things FIFO but I never did use them much.
>
>Actually, none of them is correct. Take this Josquin example,
>for instance (imagine treble clef):
     
     [ nice ascii art ]

>the a4 at (A) needs to be turned off in a stack-like (LIFO) manner, 
>whereas the f4 at (B) requires a FIFO interpretation of incoming
>note-offs.

i take it that this piece was not intended to be playable on a single
keyboard ? because ....

>Of course, since most controllers don't allow you to play a new 
>attack before the note is released (like grand pianos do), 

not strictly true. you can't play a new attack on a piano without
changing the sustain of the previous note generated by the same
key. even though you may not fully release the key, semantically
you've done something close to a noteOff before the next noteOn. 

in this specific example, its hard to imagine how you could get a
piano to generate the intended acoustic effect. or perhaps i am just
forgetting something about pianos :)

--p