I'm a bit less happy with my answer, on reflection. I'm using "complexity" to mean something close to Kolmogarov Complexity, which I still think is reasonable.

I'm a bit less happy with my answer, on reflection. I'm using "complexity" to mean something close to Kolmogarov Complexity, which I still think is reasonable. But we may need more categories to have a sensible discussion. I now begin to think we should perhaps distinguish between size, complexity of design, the number of moving parts, and I might also add redundancy (since it trades off with many of the above.) These can all vary independently. There's a sense in which redundancy (or fail-safe system design) can add parts but reduce overall design complexity, or at least management complexity/too many bad decisions that are too easy to make or implement. So maybe the latter is an additional meaning of complexity that should be considered. "Control Complexity." For example, an aircraft with no dihedral is simpler, but far harder to manage, to keep level. Note re the above that the number of moving parts (and size) in a physical system doesn't nec reflect Kolmogarov Complexity since the Turing machine simulating that system can loop.