3 ms·
> an order relation that is a partial order, not a total order. x == x (aka reflexivity) is part of the definition of a partial order[0][1] There's case to be
by a1369209993 5y ago
> an order relation that is a partial order, not a total order.
x == x (aka reflexivity) is part of the definition of a partial order[0][1]
There's case to be made that (bitwise-)distinct NaNs should compare unordered, but that's generally not what people mean when they complain about incorrect floating-point equality for NaNs.
0: https://en.wikipedia.org/wiki/Reflexive_relation https://en.wikipedia.org/wiki/Reflexive_relation
1: https://en.wikipedia.org/wiki/Partial_order https://en.wikipedia.org/wiki/Partial_order
- kmill 5y agoThe definition is just saying that "<=" is reflexive. When it says it's "a homogeneous relation on a set P that is reflexive, antisymmetric, and transitive," the meaning of P being a set is that it's an object described by set theory, and sets come with a reflexive, symmetric, and transitive equality relation -- the "native" one that comes from the underlying logical theory. (ZFC is built on a first-order logical theory with two relations: = and ∈.)
- kmill 5y ago(Sorry a1369209993, I got confused about HN threading -- somehow I thought you were responding to me.)