Nice story. But your business perspective makes you miss that scientific software from academia has always been open. We used to call it “public domain”. Lots of it was just graduate student projects, abandoned after graduation, but a lot of it lived on, sometimes with great success. To name an obvious one: Lapack out of the University of Tennessee, which lives on in Intel’s MKL and in NumPy. Scientific public domain software only started worrying about licenses when the non-scientific world invented that word “open source”.