Re: Newbie question(s)...

From: Mxsmanic (mxsmanic_at_hotmail.com)
Date: 09/17/03 

Date: Wed, 17 Sep 2003 06:02:18 +0200

M.S. Bob writes:

> If you new to this topic, why do you expect to come up with a novel
> method of encryption?

Everyone is new to the topic at some point. That is not an absolute
barrier to coming up with something new (and potentially secure).

> It is better than most first posts, but really
> why do people new to cryptography think they can whip
> off some neat novel new method that is secure, given
> that there are professionals that spend years evaluating
> new algorithms....

Because occasionally they do. New algorithms may not have anything to
do with old algorithms. While a knowledge of cryptography allows
someone to avoid mistakes, it does not allow someone to find new methods
of cryptography. If that were really possible, then cryptographic
algorithms would all fall as quickly as they were invented.

It's hard to see how RSA follows implicitly from Enigma. Clearly, a lot
of people innovated along the way, and you don't innovate empirically,
by definition.

It's intriguing to think of irrational numbers as random-number
generators. If you could find a way to generate large numbers of
irrational numbers in a computationally feasible way using a very large
number of equal seeds, you could use the result as a one-time pad
(although proving that the irrationality of the number is truly random
might be difficult). I assume that this is not computationally
feasible, or that the results are not truly random, or that the keyspace
is too small (i.e., there is no algorithm that can generate a string of
digits from an irrational number given a key of n bits with nearly n
potentially different keys). Otherwise someone would have done it.

Clearly, if there were a way to deterministically produce a string of
truly random bits based exclusively on a fixed-length key, that would be
very useful indeed. A key of n bits could generated 2^n different,
non-repeating, perpetually and completely random bit strings, so you'd
have one-time pad security, and the only attack would be a brute-force
attack against the key. With a sufficiently long key, that would be
impractical. It seems theoretically possible; the only difficulty is in
finding an algorithm that would actually do this. Or am I missing some
provable obstacle to developing such a system? 

-- 
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.

Relevant Pages

  • Re: New 2 Cryptography
    ... Newsgroups: sci.crypt ... > cryptography and I thought I'd post here to ask a few questions. ... Also there is a question of licensing one or more algorithms, ... rounds of arithmetic operations, and you wind up with something that would ...
    (sci.crypt)
  • Re: Sorting in VBA (andVB)
    ... Modifying the algorithms to handle strings as you indicate would have the ... the results using either VB 6 ListBox sort or Excel Sort. ... > order which results from a straightforward string comparison. ... > routine could then by dropped into the various string sorting algorithms ...
    (microsoft.public.word.vba.general)
  • Re: npos in string operations
    ... >>> It's related by its type to the allocator, not to the string class. ... You stated that npos should be defined by the allocator class, ... it is like a past-the-end iterator or a NULL pointer. ... > standard algorithms or other containers taking this approach. ...
    (microsoft.public.vc.stl)
  • Re: Size of RN vs USN (Was: Germany Still Loses BB...) [OFFTOPIC, BUT INTERESTING]
    ... >>That would be news to the cryptography world. ... algorithms exist which would solve these problems in polynomial time ... Sometimes it takes a lot of machines working together, ... there is this hype about quantum computes. ...
    (soc.history.war.world-war-ii)
  • Re: Normalisation
    ... The complexity I'm used to computing on algorithms deals ... >> with time or number of operations; both the set and string algorithms ... > much space will that counter take on the Turing tape. ...
    (comp.databases.theory)