Re: Solution to 7 over sqr(71) time against integer polyrhythms

[Carl Sorensen]

On 11/17/16 7:21 PM, "mclaren" <address@hidden> wrote:
>This kind of trolling by Kieren (and others) is not useful.  Yes, you can
>bend and twist Lilypond into printing out tuplet numbers that bear no
>relation to the actual tuplets generated by Lilypond...but that's not what
>was asked for. Any reasonable person understands that the point of this
>entire example is actually to get Lilypond to generate those tuplets so
>that
>Lilypond can produce the MIDI for a Nancarrow-typle 1% acceleration,
>meaning
>nested tuplets in the ratio 100:99, and then Lilypond can print out the
>actual score and not a fake score done by inserting numbers on the page
>that
>are never actually used in the music.

I agree with part of what you say here, but not all of it.

The part I agree with is that Kieren's code (using 10:9 tuplets but
printing 100:99 tuplets) is a hack in that the printed output does not
match the semantics of the musical input.  This is clearly less than ideal.

The part I disagree with is the implication that the problem is that then
LilyPond won't create the right MIDI.  LilyPond is aimed at creating
printed output, not MIDI output. MIDI output is a side benefit, not the
primary reason for being.

So while it would be better to be able to actually use 100:99 nested
tuplets (and I hope that you can figure out how to do it), for the primary
purpose of LilyPond (which is producing printed scores), it would be quite
straightforward to make a music function that would fake the tuplets.

As I said, it's not optimal.  But it's considerably better than using
photoshop or inkscape to create the score.

And I'm sure that we really would welcome a patch that solved the problem,
as long as it didn't hugely slow down processing of more traditional (e.g.
19th century) music.

Carl