Updating the Historic OTP.
in [Cryptography]
|
Prev: A C-code for permutation polynomials mod 2^n
Next: On selection of polynomials for the Multiple Polynomial Quadratic Sieve
From: Bruce Stephens on 25 Mar 2010 06:15 WTShaw <lurens1(a)gmail.com> writes: > On Mar 24, 4:01 pm, Bruce Stephens <bruce+use...(a)cenderis.demon.co.uk> > wrote: >> Mark Murray <w.h.o...(a)example.com> writes: >> > On 24/03/2010 08:01, adacrypt wrote: >> >> characters numbered 32 to 126 inclusive => 95 characters >> >> > Why? >> >> Apart from anything else it means you're leaking paragraph lengths (and >> likely line-lengths, presuming one uses lines). That's the kind of >> thing that you definitely shouldn't be doing. >> >> [...] > > Given that information, which could be wrong on purpose, solving it > might be a dare to try to illustrate the strength of a system. It's an obvious weakness that wouldn't be present in a straightforward OTP implementation: it's one that has been *added*. Of course you could try to hide newlines. You might just not use any (and not use paragraphs). Or use base64, for that matter. But why bother? Why not just do it sanely? (As described on p.16 of Applied Cryptography, for example---why didn't adacrypt just turn the page?)
From: David Eather on 25 Mar 2010 06:49 On 25/03/2010 4:01 PM, WTShaw wrote: > On Mar 24, 4:01 pm, Bruce Stephens<bruce+use...(a)cenderis.demon.co.uk> > wrote: >> Mark Murray<w.h.o...(a)example.com> writes: >>> On 24/03/2010 08:01, adacrypt wrote: >>>> characters numbered 32 to 126 inclusive => 95 characters >> >>> Why? >> >> Apart from anything else it means you're leaking paragraph lengths (and >> likely line-lengths, presuming one uses lines). That's the kind of >> thing that you definitely shouldn't be doing. >> >> [...] > > Given that information, which could be wrong on purpose, solving it > might be a dare to try to illustrate the strength of a system. You know what those federal agencies think of that type of "dare" - it's just fuel to fire up "old sparkie"
From: Mark Murray on 25 Mar 2010 17:00 On 25/03/2010 08:08, Boon wrote: > What is it about sci(ENTIFIC).crypt(OLOGY) that draws so many > kooks, cranks, and non-understanders? Nothing specific. There are cranks in every discipline. And not only that, sports, crafts, arts etc. There is no limit to the spaces available where an individual may overestimate him[*]self for the amusement of others. I have first-hand experience of physics cranks when working in my first "real" job in a university physics dept. I've been scared witless by a skydiving crank (I have 400 jumps myself). My mum (very "arty", and an retired teacher) has seen the lunacy of Picasso-wannabes. Ain't nothing special about sci.crypt. Its just an easy target. M -- Mark "No Nickname" Murray Notable nebbish, extreme generalist. [*] I have no memory of a female crank, but then I have a non-exhaustive sample and a lousy memory.
From: WTShaw on 26 Mar 2010 16:56 On Mar 25, 3:00 pm, Mark Murray <w.h.o...(a)example.com> wrote: > > I've been scared witless by a skydiving crank (I have 400 jumps > myself). > Wantabe test pilot?
From: WTShaw on 26 Mar 2010 17:00
On Mar 25, 4:15 am, Bruce Stephens <bruce+use...(a)cenderis.demon.co.uk> wrote: > WTShaw <lure...(a)gmail.com> writes: > > On Mar 24, 4:01 pm, Bruce Stephens <bruce+use...(a)cenderis.demon.co.uk> > > wrote: > > > > Given that information, which could be wrong on purpose, solving it > > might be a dare to try to illustrate the strength of a system. > > It's an obvious weakness that wouldn't be present in a straightforward > OTP implementation: it's one that has been *added*. > Obvious and oblivious are not the same. |