It's not used or required. But it's common to only talk about $G/H$ when $H$ is a closed subgroup, because otherwise $G/H$ will not be Hausdorff. In particular, the book in which that proof appears defines topological groups to be Hausdorff (see page 84).
It's not used or required. But it's common to only talk about $G/H$ when $H$ is a closed subgroup, because otherwise $G/H$ will not be Hausdorff. In particular, the book in which that proof appears defines topological groups to be Hausdorff (see page 84).