# When is $[0,1]^K$ submetrizable or even metrizable?

Let $I=[0,1]$ and $K$ is a compact space. Then could the function space $I^K$ be submetrizable, even metrizable? In other words, in general, if $I^A$ can be submetrizable (metrizable) for some space $A$, what's condition that $A$ should satisfying?

-
What topology do you have on $I^A$? If $A$ is uncountable, the product topology is not metrizable. –  Michael Greinecker Dec 30 '11 at 8:29
Really, I want to know the results on the topology of uniform convergence, the topology of pointwise convergence and the compact-open topology. –  Paul Dec 31 '11 at 2:56

If $A$ is compact, $I^A$ is metrizable with the metric being the uniform norm. That is, $d(f,g):=\sup_{a\in A} d(f(a),g(a))$.
Do you mean that if $A$ only needs to be compact then $I^A$ has this metric $d$? –  Paul Dec 30 '11 at 2:44
@John Correct. In general, $X^Y$ is metrizable if $Y$ is compact and $X$ is a metric space. Hatcher's Appendix has a proof of this (Proposition A.13). –  SL2 Dec 30 '11 at 3:03
It's nice! Do you know other conditions on $A$ such that will make the space $I^A$ metrizable? For example $A$ is countable and so on. –  Paul Dec 30 '11 at 3:19
$A$ sigma-compact and locally compact Hausdorff will also guarantee that all functions from $A$ to $\mathbb{R}$ in the compact-open topology will be metrizable, and this is necessary as well, for functions with codomain the reals. –  Henno Brandsma Dec 30 '11 at 12:42