Re: [Monotone-devel] Re: current multiple heads (was Re: write access to my public server)

[Nathaniel Smith]

On Fri, Apr 30, 2004 at 08:43:16AM +0200, Jon Bright wrote:
> graydon hoare wrote:
> 
> >the controlled method is very simple: a file called .mt-nonce goes in 
> >the working copy which contains some respectably large number (say 4096) 
> >of random bits. a new command -- "monotone bump" -- generates a new set 
> ...
> >I won't be implementing this in time for 0.12, but it strikes me as 
> >quite feasible within the 0.13 timeframe. does anyone object? it's a 
> >kludge, but a simple and obvious one.
> 
> Are these cryptographically random bits?  If so, using that many random 
> bits does sound like a bit of a recipe for draining the machine's 
> entropy pool (particularly on unattended remote machines that don't have 
> the advantage of keyboard/mouse input).

4096 bits is a pretty absurd number of bits.  GUIDs are only 128 (and
they waste some of those), and these don't have to globally unique,
just unique among each reverted version of a single tree.  We could
probably get away with only 32 bits.  For that matter, it doesn't
really have to be cryptographic entropy in the first place; you're
going to be sharing it with everyone who matters anyway.

> Would generating a GUID perhaps be easier?  It'd certainly be less 
> entropy-hungry.  And there's lots of pre-existing stuff about for 
> dealing with GUIDs...

Well, if we used the sort of GUID that's generated randomly, that
would be similar :-).  The other sort of GUID is tricky to do
portably (POSIX doesn't provide a way to get the ethernet card's MAC
address, alas, especially on computers without ethernet cards), and I
don't know that they would work well with XOR merging, since there's
a lot of static stuff in them.  (Like MAC addresses, for instance.)

-- Nathaniel

-- 
Sentience can be such a burden.