The difference between {\x >\y} and {\x >? \y} in xfp package?

Question

On page 1/2 of xfp package we see x >? y as a valid comparison operator. Is it different from x>y? I think they are equivalent based on my following code, but not sure:

  \documentclass{article}
\usepackage{,xfp}
\begin{document}
\edef\x{5}
\edef\y{6}
\fpeval{\x>? \y}\
\fpeval{\x> \y}
\end{document}

The result for both is zero. So are they equivalent in general?

Answer 1

The question mark ? is the “not ordered” comparison of l3fp. More or less quoting from the documentation of \fp_compare_p:nNn:

x ? y is true if x and y are “not ordered” which occurs exactly if one or both operands is nan or is a tuple. Note that nan is distinct from any other value, even from itself, i.e. if x = nan then x == x is false. To test whether a value is nan you can use the “not ordered” comparison with any other value, e.g. x ? 0.

That also means that if neither x nor y are nan or a tuple, the comparison x >? y will be equivalent to x > y.

---

Personal note: I'm rather skeptical of the usefulness of this “not ordered” comparison. It seems to me that this is trying to solve a problem that is already covered by other operators. For example in a traditional programming language, like C, you can check for nan by checking whether a value is equal to itself, i.e. x == x will only ever be false, if x is nan. For comparison between tuples, all operators should just return false (or better throw a type error), except for == and != which ought to perform and element-wise comparison.