Github for science sounds like a great plan? But do we have the underlying stack of equivalent services needed to provide “http for science” and “tcp/ip” for science. I argue that until we do we will struggle to really deliver on the excitement that examples (rightly) inspire.