Based on my own experience from
my experiments with using fractals for composing i think it's best if you go for something like this:
Let the user configure which notes are allowed for which instrument or midichannel and then try to work out how you will let the results from the fractal calculation translate which notes to pick. Also an option to set a few notes from a single octave and then an option to choose which octaves are allowed for the user selected notes. I would suggest implementing various techniques for translating the results into notes.
Also i think you could try to implement to use ALL results of the equation or to leave the equation running and select notes at intervals from the running equations.
And ofcourse let the user put in some location settings and iteration counts etc.
You need to implement a quantizaton option and settings aswell to force the note selection into a beatmeasure , but also leave in the possibily to leave it running without quantization.
You should see things like volume, panning and filter effects as the "tonecolours" of the equation results, but users must be allowed to set minimum and maximum values to prevent too radiacal changes.
Perhaps it's best to use seperate (user selectable) equations for notes and for "tonecolours".
I can't help you with any of the math nor the programming, but i think i have a pretty good idea on how to help turn this into something usefull. i have quite a few more ideas and i would like to help out in any way i can. I would very much like to do some experiments in fractal composition again.
I hope this was of some help for starters.