If we consider the set of all straight lines in $\mathbb{R}^d$, then what is it topologically? If it's topologically 'nice' i.e. a manifold, probably we could put a smooth manifold structure on it too. Can we put a Riemannian manifold structure?
For the first question, I'd think of it as a disjoint union over the points $p$ of all lines passing through a fixed point $p$. So I'm thinking it's something like $\mathbb{RP}^{d-1}\times \mathbb{R}^d$. But is it? If it is, I guess the other questions might be answered from it.
Let $L^n$ denote the set of lines in $\mathbb R^n$. It's not the disjoint union over $p$ of lines passing through $p$, because for $p\ne q$, the sets $\{\text{lines passing through $p$}\}$ and $\{\text{lines passing through $q$}\}$ intersect. It's also not equal to the Grassmannian $G(2,\mathbb R^{d+1})$, because it's not compact.
One way to topologize $L^n$ is as a vector bundle over $\mathbb R\mathbb P^{n-1}$. The fiber over each line $[x]\in \mathbb R\mathbb P^{n-1}$ is the set of all lines parallel to $[x]$, which obtains a vector space structure by viewing it as the quotient space $\mathbb R^n/\operatorname{span}(x)$. The projection maps each line $\ell$ in $\mathbb R^n$ to the unique line parallel to $\ell$ and passing through the origin.
It obtains a smooth structure as a homogeneous space: the affine group $\mathbb R^n \rtimes \operatorname{GL}(n,\mathbb R)$ acts transitively on $L^n$, with closed isotropy group $\mathbb R^{n-1}\rtimes \operatorname{GL}(n-1,\mathbb R)$.