Re: Issue 687 in lilypond: Enhancement: inequal MIDI quantization of equal durations (swing, rubato)

[lilypond]

Comment #12 on issue 687 by arvidgr: Enhancement: inequal MIDI quantization  
of equal durations (swing, rubato)
http://code.google.com/p/lilypond/issues/detail?id=687

This is perhaps a silly idea, but should we really limit this to MIDI?

Personally, I'd like a 2:1 swing { c8 c8 } to be printed, as well as  
performed, as { c8*4/3 c8*2/3 }.  If I have a triplet in a different voice  
(again assuming 2:1 swing), that means I'd want the second swing eight  
aligned vertically with the third note of the triplet.

(I don't really know if this is usual practice or not.  As far as I could  
look up, Read doesn't mention this at all.  But in my opinion, it would  
make quite a bit of sense.)

This is an implementation strategy that should be feasible to do in  
Scheme.  (Doing this manually, with 8*4/3 and 8*2/3 (and 4.*10/9 and...)  
all over the place, is certainly possible in Lilypond input.  And equally  
certainly tedious.)

Of course, anyone wanting swing in MIDI only could have different score  
blocks for printed and MIDI output, doing

  \swing #2 #1 \themusic % or whatever

...only on the MIDI part.

To re-iterate, I'm proposing a music-function (swing a b sequential-music)  
(integer? integer? ly:music?), possibly based on a more generic framework,  
where:

  \swing #2 #1 { c4. c8 c8 c8 }

gives the same as the following input:

  { c4.*10/9 c8*2/3 c8*4/3 c8*2/3 }

A possible more generic interface could be something like this:

  (define-music-function (unequal duration weight-list sequential-music)
     (ly:duration? list? ly:music?)
     [...]

Anyway, this approach does not touch MIDI at all.  It all happens very long  
before the music touches the C++ MIDI code.  It should even be possible to  
implement this as a preprocessor to Lilypond.  In Perl.  Stupid, but  
possible.

So is there any reason this can't be done in Scheme?