This is a tiny proof-of-concept banged out in an hour after a fun conversation
with Mr. Justin George ( http://github.com/jaggederest ), who had the idea that
an ordered list of words could function much better than a SHA1 (in terms of
human memory, anyway) but could hold the same information.  Sort of a look-up,
with each word representing n-bits.  It wouldn't need to be a SHA1, though; it
could be arbitrary data.  The initial idea, though, was that if you have a
reliable mapping and you make sure that the reverse-mapping works across
machines and versions of the library, then it could be a much easier to remember
way to identify git commits.  Could also work for IP addresses, SSH-keys, etc.,
so we're playing around to see if it's workable.