Here are ten random results from my first mountain test. I think there's some good there, and some not-so-good. Without modeling plate tectonics we have no guarantee that mountains will always be realistic, so the question is: are they "good enough"? Does every world need to be a winner, or might typical usage involve generating a few worlds and seeing which one you like best? Program parameters could include how clumped-together the landmasses are, how frequent and how large the mountain ranges are, what the sea level is, and so on, so people would be able to tune the output.