In case of an approximation/cp-decreasing algorithm it might be a good idea to start open source. I agree. As I said if I find time and motivation I will start such project, because I really believe in the use for Rhino and I already coded some variants.
Generally seen its a more difficult question. And a dilemma.
Because the big drawback of open source licensing is that practically nobody cares about different types of free licenses as long as if its not obvious stealing. But even then, how would you be able to enforce your right if the thief is sitting in another country, or if they are too powerful? In Front of a court, you are nothing more as sailor on a stormy sea. Your fade is in “gods” hands.
So I always have seen obfuscated c# as something in between. Allowing fast development, but with a basic protection. My personal interest is nothing more as creating “opportunities”, maybe getting invited somewhere, meeting other people. Nobody truly works for free. Everyone has some intention; monetary or not. But as I pointed out in my other post about worklife-balance. I really ask myself if open source really leads to any personal benefit. Because if people can simply copy and paste, they don’t care about you anymore.However if you provide a free-to-use software, but encrypted code it might be different. Since people need you for further or specific development.