+1 for the idea of a module for matrices and vectors in the stdlib. One aspect that I would like to highlight is that NumPy is not always so easily available. I got feedback from many schools that they were using tablet computers or something like Chrome books and thus basically limited to anything available in the browser. Projekts like Skulpt have done a tremendous deal here to make Python available even under such restricted circumstances. However, porting the entire NumPy library to an "alternative" implementation is a huge project and usually well beyond the possible. A simple module for vectors and matrices, on the other hand, might be readily portable and thus make a big difference! Vectors and matrices are also very versatile tools for a lot of applications. I have actually seen quite a number of programs that start all by defining some sort of vector class (particularly in graphics, as Greg Ewing pointed out). But as a physics teacher, say, you usually simply don't have time to start defining vectors and matrices in your course, but just want to use them. And if Python came out of the box with it, that would in turn help make a point for learning Python at schools. Long story short: I fully support the idea and think there are valid use cases where NumPy is not really an option (for whatever reason).