The problem there is using uninformative names, rather than URLs/URIs. Using the latter would be quite nice, since (as a reader) we could click on any of these

The problem there is using uninformative names, rather than URLs/URIs. Using the latter would be quite nice, since (as a reader) we could click on any of these links, or spider them to download the whole transitive closure of prior results; as a developer we could query those links to show e.g. their latest name, whether they've since been retracted/disproved, etc.