That would be excellent to hear about useful forgotten patches.
In the meantime maybe I can hijack this thread with a practical question. Look at the supercollider and pure data example at the beginning of the following tutorial:
http://en.flossmanuals.net/pure-data/ch002_graphical-programming/
Now it's perfectly fine that it shows one example that gives the strengths of supercollider and a completely different example for the strengths of pure data. But it certainly begs the question what each example would look like in the other enviroment.
So: what's the simplest way to implement that supercollider example above using pure data? Using nqpoly? Dynamic patching? Is there a #many~ in GF? Etc.
-Jonathan