Re: [Monotone-devel] Proposal for human readable revision IDs

[Andy Jones]

Well, yes, it is just another sort of certificate.  But you would need
to change the monotone code in order to make proper use of it - for
example, monotone ls revisions would want to output it; you'd want to
be able to select by it.  So you can't do it as the code stands.  What
I mean is, there would be no point in having a "nickname" certificate
if it wasn't used in the monotone code.

Regarding uniqueness, I'm not bothered.  I don't think nicknames need
to be unique; but if you use them for selection, Monotone should
detect nickname collisions and complain.

If you want a guaranteed unique name for the revision, use the
revision hash id, as now.  The same applies to the "abbreviated
revision ID" attack outlined above: you'd want to use the revision ID
in your 'montone pull' to make sure you got the right one; but from
then on you could use the nickname.