Re: M4 project

On Tue, 2006-03-07 at 16:32 +0000, Geoff Sullivan wrote:
> steve-ZMQM5VTtYHJ+cjeuK/[email protected] wrote:
> > 
> > I'm contributing to the M4 project, but am a bit surprised at the
> > tactics. I don't have a problem with the *concept* as such, but with the
> > bit that says "lock the middle wheel ring at 'A'".
> 
> If M4, 40% chance there is no double notch wheel. 60% chance with Service wheels that there is no slow turn-over. Therefore the tactic is to try the easier options first.  If that fails then try different rings. 
> 

Yeah, but they apparently did *10* walks through the whole M4 keyspace
with the middle ring at 'A' if I understood it right. Surely if the
first walk failed, there's no point in doing it again? And certainly not
without changing the mid wheel ring.

> 
> > no slower wheel to use its notch. The fast wheel ring is dealt-with as
> > it must be - wrong ring there is fatal.
> 
> Wrong fast ring is not always fatal. See todays break. 

OK. Apart from middle ring just 1 or two letters out either way, I'd
have thought the garbles would break the stats. Evidently not.

Presumably one of the optimisations in the system is to try only every 5
fast wheel ring positions on the grounds that you'll hillclimb OK even
if you slightly miss the correct ring position.

> > And why do M3 at all? With the M4 fourth wheel at 'A' it emulates an M3
> > anyway! That was the secret of the short weather cipher crib after all.
> 
> I thought M3 (equivalent to Greek at A) was done first, This setting would not be included in a subsequent M4 run. Although I'm not involved in the project so don't know all the details.
> 

Yeah, but there's no need AFAICS to special-case M3. You'll do M3 for
free on those M4 setting where thin rotor beta is at 'A'.

Was there even a ring on the thin rotors at all?
 
-- 
Steve Hosgood <steve-0L79qhCdni9Wk0Htik3J/[email protected]>