[PD] permutations

[Orm Finnendahl]

hi Krzyzstof,

more of that brainfood, please! It's really permutaining. Apart from
being an interesting approach, it taught me a lot of things about pd
(and its differences to MAX), I didn't know. There should be more
examples out there which show different programming styles.

About being generic: If the english meaning of "permutation" is the
same as the german, permutation means a mapping of positions as
implemented in my previous patch. What you proposed we call
"combination": It refers to all possible combinations of a certain set
of numbers (maybe even with a given subset, like in lotteries, where
you select 6 numbers out of 49). There are lots of interesting
properties of permutations, regarding their periods etc. which make
them a somehow special subset of all possible combinations and that's
probably the reason, why especially composers in the 50s really liked
them (otherwise the pieces would have gotten even longer...).

Among the externals, I wrote for Max was also a patch, which does all
combinations (of sets and of subsets of sets). This external has the
advantage of being indexable and of telling you the exact number of
possible combinations. I don't remember the exact approach about the
systematics anymore, but I know, I had to make it up myself, too. I
remember not being too satisfied with it. I'll send the original
C-code to you in a seperate email together with the code for the
permutation external. That way you can find out, how it was done.

Anybody else interested in it should let me know.

--
Orm


P.S.: I have to admit that I prefer a text based approach to solving
problems like this, as I find it more comprehensible to read

  for (i=max;((i > 0)&&(!found-first)); i--) ...

than to clutter the patch with int and +1 and trigger and until boxes
just in order to implement a simple loop. That's probably, why I made
code externals and after a while switched to James Mc'Cartneys
"pyrite" object for Max (some sort of programming language with
embedded text editor in a max object) before more or less ending the
Max business altogether some 6 years ago. But that's probably poison
for a pd-list =:-)

Am Donnerstag, den 24. Januar 2002 um 21:01:29 Uhr (+0100) schrieb Krzysztof Czaja:
> hi,
> 
> since Orm's [pd permute] cycles with a period of 6 (out of 5! == 120),
> I have just patched a correct(?) one as an evening refresher, trying
> to make it a bit more generic too...
> 


> This one is clumsy!  But is there much simpler way?
> 


> Or is this yet another case of my blindness?
> 
> Krzysztof