Quote:
Originally Posted by checktheriver
It's a strange problem...I'm pretty sure you can't find a basis for U as a vector space without appealing to the axiom of choice.
Yeah for someone who hasn't talked much about bases this is a strange problem.
If the problem were to find a basis for V, a subset of Abb(R,R) such that each element of V is zero except at a finite number of points in R, then a basis would be {f_a(x): a in R}, where f_a(x) is 1 at x=a and 0 elsewhere.
But this is not a basis for U, since we can form every element of a space by a
finite linear combination of basis elements. And elements of U are allowed to be nonzero at infinitely many points in R (though a set of measure zero. Take f(x) = 1 if x is an integer, 0 otherwise, for example).