In fact, this shows that “when optimizing code, know what your asymptotes are”. It is more than an order of magnitude trivial difference between 34 and 1.2, but the conclusion bases itself on silly 0.1 which plays no sensible role at all.
People take advices like “don’t optimize” and then in 2030 someone 300byte-patches system updates so they would install at ~ssd write speed instead of sudden 1.5 hours. And everyone is like wow, how we overlooked that. You never hired an engineer, that’s how.
Your comment shows that you've missed the point. You're assuming that because I'm doing this kind of micro-optimisation that is, in the grand scheme of things, irrelevant, that I'm concentrating on the wrong thing.
You say:
> the conclusion bases itself on silly 0.1 which plays no sensible role at all.
You've missed the point. This specific code change is not the point for the task overall. This specific code change was preparation for a significant algorithmic shift that was about to happen. On the way I happened to notice something odd, and I thought people would be mildly curious, so I wrote it down for people to see.
Please don't assume that I'm clueless about the wider picture.
People take advices like “don’t optimize” and then in 2030 someone 300byte-patches system updates so they would install at ~ssd write speed instead of sudden 1.5 hours. And everyone is like wow, how we overlooked that. You never hired an engineer, that’s how.