Comment by leoh

I couldn't agree more strongly. There is a huge opportunity to make git more effective for this kind of use-case, not to abandon it. The essay in question provides no compelling alternative; it therefore reaches an entirely half-baked conclusion.