Dec. 29, 2019
11:19 p.m.
On Sun, Dec 29, 2019, 5:20 PM Andrew Barnert via Python-ideas
But it is, out of all of the possible magma-over-magma structures on those values, the one that most closely approximates—in a well-defined and useful, if very complicated, way—the rationals.
I'm sort of convinced that Posits better approximate the behavior of rationals than do IEEE-754 floats:
https://en.m.wikipedia.org/wiki/Unum_(number_format) But that's a small footnote. Until or unless this system enters widespread use, tradition floats are our system. And anyway, all the philosophy of mathematics stuff I wrote before would apply to Posits equally, just with somewhat different distributions of errors.