hash algorithm names
Apparently
"hash algorithm names" aren't necessarily all written the same everywhere. Here we say "sha3-256", but I can't guarantee that we don't elsewhere call it "SHA3_256" or something.
We should have a single registry of hash functions, where each function has one textual name for use in places like netdocs, and one numerical id, for use in binary protocols.
That might involve breaking changes. On the way towards sanity, we will probably need to inventory what the current situation is.