> Another problem with seeing purity as fundamental is that our physical reality isn't a persistent data structure (past states aren't accessible), so the best

> Another problem with seeing purity as fundamental is that our physical reality isn't a persistent data structure (past states aren't accessible), so the best algorithms possible won't treat it as one. Among the tons of papers describing fast algorithms, practically none are using purity, unless they required purity to begin with. Purity is a simplification of the real world, in the same way that the abstract ideal of a circle is a simplification of objects in the real world. But that simplification yields great benefits (kind of in the same way that ideal geometric shapes allow mathematical reasoning). It seems to describe the simplifications that the real world "approaches" (like a mathematical limit). So it does seem fundamental. But yet not "real." I don't think those are contradictory. It makes intuitive sense that non-purity could open up algorithmic possibilities, some of which would be faster. But that doesn't really touch on whether purity is "fundamental" in the sense described above.