So I got a wild idea to try and do some of Structure and Interpretation of Computer Programs, but in PD rather than scheme. At "1.3.1 - Procedures as Arguments", there's a bit of code that goes

(define (sum-integers a b)
  (if (> a b)
      0
      (+ a (sum-integers (+ a 1) b))))

that sums integers a thru b. I tried to make a patch that does this, it is attached. I was super stoked, and think it is correct, but PD won't run it, says "can't load abstraction within itself".

That's sad to me. I'm reading some stuff about distributed and parallel algorithms, and could really see myself reading a lot about that stuff and really getting into it. I'd really like to see it allow subpatches and custom objects to contain instances of themselves.

I just saw this patch 50.somethingsomething.pd in the extended release that creates a clone of itself! wicked.

now that I'm looking at that sumints patch more I think its kinda wrong... oh well. more staring and thinking.

Any thoughts on these issues would be much appreciated. Haven't posted in a while, good to look around once more! -walt

http://www.pdpatchrepo.info/hurleur/sumints.PNG

it's not trivial, but it is possible with some dynamic patching.
-see attached patch

honestly though, i think you should look more into creating loops with [until] or something like that if you want to do recursion in pd.

http://www.pdpatchrepo.info/hurleur/recur-sum.zip

The problem with recursion in Pd has to do with the how the functions are called in typical line code and how abstractions are called in Pd. Functions are only called when explicitly asked for when the program is running. I don't know Scheme, but it looks like that only calls the function if a<=b. Pd, on the other hand, calls abstractions when they are instantiated. Thus, if you put an abstraction within itself, you'll have infinite recursion.

You can do recursive-like things without dynamic patching (though I don't know if it's the case for all purposes) by simply reusing the abstraction itself. You basically check to see if the condition is met and, if not, send the data back into it. It's a little late, and this might not be the best patch to illustrate, but I attached one that should at least give you an idea what I mean.

mod's right, though. For this particular situation, [until] is easier, but I'm guessing this is for teaching purposes. I'm sure Scheme has some sort of for...loop that would also be easier.

http://www.pdpatchrepo.info/hurleur/sum-integers.mmb.pd