open source capital.
Basically, we phase in object-oriented talk when introducing such concepts as vectors and polyhedra, and then we use those vectors to actually build those polyhedra on screen. I didn't go in to all the details, although I did hold up the MITE cube and dissect it (this was a video linkup).
In Germany, there's more of a top-down, "government tells us" approach, whereas the USA has this tradition of treating states as laboratories, where many competing approaches run in parallel, a kind of genetic algorithm wherein new hybrids at least have a fighting chance of finding a place in the sun.
Our computer math hybrid starts using Python pretty early, mainly to generate sequences at first, such as the Fibonacci numbers and successive rows of Pascal's triangle (e.g. see Pippy on the XO for a currently non-generator-based approach).
However, we also mix in more semi- and non-numeric algorithms (e.g. around XML), plus some elementary group theory with an eye towards explaining RSA in some detail by senior year high school (group theory also helps in later physics, intimately connects to those polyhedra and their rotations).