Not for homework, I am trying to self study functional analysis and encountered the following problem.
We let $C[0,1/2]$ the continous functions defined on that subset of the real line. We look at a subspace of $C[0,1/2]$, consisting of all polynomials on $[0,1/2]$ and call it $W$. Given $\delta>0$, set.
$g(x)=\delta\sum_{n=1}^{\infty}\frac{1}{n+1}x^n$
Where $x\in[0,1/2]$. First, we are tasked with showing that $g$ is in the open ball $B(0,\delta)$. I suppose this is done by showing that the norm of $g$ must be less than delta (we are using the supremum norm inherited from $C[0,1/2]$). But from that result, we are tasked with using it to conclude that $W$ cannot be an open subset of $C[0,1/2]$. So I must show that for some $w\in W$, there is no $\epsilon$, such that an open ball $B(v,\epsilon)$ is in $W$? Not sure how that follows from the previous?
$\|g\| < \delta \sum_{n=1}^{\infty} \frac 1 {2^{n}}= \delta$. This proves the first part. Now,suppose $W$ is open . Since the zero polynomial is in $W$ there must exist $\delta >0$ such that $B(0,\delta) \subset W$. Consider the $g$ corresponding to this $\delta$. Then $g \in B(0,\delta)$ so we must have $g \in W$. Can you see that this is a contradiction? [It is a known fact that if $\sum a_n x^{n}$ converges for $|x| \leq r$ and if the sum is zero for all such $x$ the $a_n=0$ for all $n$].