Re: SPES (my new encryption) one of its kind



do you think using higher number of ciphers really cost the user more
than the hacker?
No, in this I qiute agree with you (in fact my work Kyu is higly
focused on cypher composition, the framework leave room up to 16
syncronous cyphers, with some of those being adaptative driven by the
key and or the message... quite a lot!) and it's simple to calculate
it: bruteforcing an unknown composition of n layers of encryption
between a m cyphers wide pool would cost at the attacker m^n
computational cost while to the receiver it will cost only n, so in
that sense this would be reasonable, hoever it's very difficult to
understand and prove anything about that construction.
Instead, with a cleaner construction of a simpler cryptosystem we may
be able to understand and prove desired characteristic that may lead us
to trust that algorithm up to a given keysize, and increasing the
trustable keysize is even a more efficient way to act because it will
increase exponentially the bruteforcing effort but only marginally the
user computetional workload.
And, as I said, it's also important to build a system with a lot of
useful characteristic, not a single overkilling characteristic that as
an armored door for an house with open windows doesn't give a single
more chance to resist to an attack, that would invariabli be led toward
weaker points (cycle lenght? key schedule? weak keys? etc...)

.