I think we can view paradigm shifts in science as a gradual accumulation of evidence (data), leading to more sophisticated and complex models until someone sees

I think we can view paradigm shifts in science as a gradual accumulation of evidence (data), leading to more sophisticated and complex models until someone sees a new way of thinking about it that explains the evidence more simply. The evidence accumulated before the shift is essential. In programming, we're not seeking theories, but solutions to problems, so I think the analogy is that as we try to solve problems, we learn more about them (collect data), creating increasingly elaborate solutions, until eventually someone realizes there's a better approach. Again, the accumulated knowledge about the problem is essential for the breakthrough. In science, reality is constant and it's only our knowledge and theories that change. But in programming, the actual problems also change as we enter new domains with new problems and criteria.