from Hacker News

Universally Unique Lexicographically Sortable Identifier

by bound008 on 5/12/20, 10:11 PM with 2 comments

• by some_furry on 5/12/20, 10:29 PM

  > Uses Crockford's base32 for better efficiency and readability (5 bits per character)

  From what I can tell, this is the alphabet:

    0123456789ABCDEFGHJKMNPQRSTVWXYZ
  

  Is there any reason why Crockford's alphabet hasn't been rolled into an RFC that supersedes RFC 4648 if it's more efficient and readable?

  RFC 4648 specifies [0-9a-v] (base32hex) and [a-z2-7] (base32).

• by rurban on 5/13/20, 3:49 AM

  A long uppercase trash is hardly considered a recognizable identifier. Remember, it needs to be identifiable. IMHO UUID's are more identifiable with its 4 words.

  Uniqueness is usually guaranteed by a symbol table, and the op is gensym().