Suggested by @jturner314:
I'd just do it in terms of fold with something like this (taking advantage of the first method from PR #507):

impl<A, S, D> ArrayBase<S, D>
where
    S: Data<Elem = A>,
    D: Dimension,
{
    /// Returns the minimum element, or `None` if the array is empty.
    fn scalar_min(&self) -> Option<&A>
    where
        A: Ord,
    {
        let first = self.first()?;
        Some(self.fold(first, |acc, x| acc.min(x)))
    }
}

We don't need to manually unroll this because the compiler does a good job automatically (checked with Compiler Explorer using the -O compiler option).

The desired behavior for floating-point types depends on the use-case because of NaN. One option is

arr.fold(::std::f64::NAN, |acc, &x| acc.min(x))

which ignores NaN values. (It returns NaN only if there are no non-NaN values.) The compiler does a decent job automatically unrolling this, so we don't need to manually unroll in this case either.

Originally posted by @jturner314 in #505 (comment)