I have contracts which use name as the type of the primary key in several of their tables. This exact use case is historically the cause for the name type to have been envisioned and created in the first place. Even the EOSIO core contracts made extensive use of this pattern, albeit, by bending over backward to protect the CDT templates from the true type of the key...

Well that's just a bunch of silliness! Updating the templates to handle name typed primary keys is trivial, so why not just solve the problem once and have it go away, rather than adding complexity to every other contract?