More generally, only very rarely should canonical identifiers for computer use ever be meaningful to humans. The (potential, and conventional and UI-encouraged)

More generally, only very rarely should canonical identifiers for computer use ever be meaningful to humans. The (potential, and conventional and UI-encouraged) meaningfulness and tree structure of the HTTP URL path component has probably contributed seriously to the Web's link-rot. Having a tree-structured, human-readable name at the top of the browser window is great! It just shouldn't be the URL.