How do I convince my researcher colleagues to release their code, no matter how badly written they think it is, for others to see?