Re: Using MGF1 for key generation
Date: Tue, 02 Aug 2005 21:49:09 +1000
Tom St Denis wrote:
> Crypto@S.M.S wrote:
>>Joseph Ashwood wrote:
>>>The only downside is that as Paul Rubin points out there are
>>>reasons to use SHA-1 (or 3DES) at this point which is why I always recommend
>>>that everyone prepare to rollover to a different cipher/hash/etc when
>>How often is it necessary to rollover to a different cipher/hash/etc?
>>Is this required rollover due solely to the failure of a particular
>>algorithm, and if so, would it not extend the time between failures
>>to use more than one cipher/hash/etc?
> No you ignorant lazy crank, not every situation can call for wasting
> cpu cycles, power or memory to implement a half dozen different
It is not a "waste" of CPU cycles to protect against future failures.
How it is lazy to implement a half dozen different algorithms? You
seem to be confusing extra work with sloth.
Every situation, that calls for security against future attacks, calls
for as many algorithms as possible. It is laziness to presume that
one single algorithm will be perfectly strong forever. Lazy,
short-sighted, and ignorant of cryptographic history.
> How about you write cryptosystems for an 8051 and then come back to us
> with the "glorious merits of multiple encryption". Even the PocketPC
> is too powerful [and your desired efficiency too low] to show off how
> bad your "design is".
I have written 1000's of lines of code for embedded 8051 systems, mostly
medical life support equipment, but also quite a bit of firmware for
And if the CryptoSMS design is so bad, please demonstrate a practical
weakness. So far you have offered nothing but insults and nonsense, and
then call me a crank for pointing that out.
> Go out and get some real world experience.
Already have plenty. Been writing code for more than 30 years.
Wrote my first assembly language on an old PDP-8.
Again, please stick to the facts and quit trying to insult someone
you know nothing about.