[43313] in Cypherpunks
Re: Who needs time vaults anyway?
daemon@ATHENA.MIT.EDU (Bill Stewart)
Mon Nov 13 01:18:40 1995
Date: Sun, 12 Nov 1995 22:13:08 -0800
To: Simon Spero <ses@tipper.oit.unc.edu>
From: Bill Stewart <stewarts@ix.netcom.com>
Cc: cypherpunks@toad.com
At 01:15 PM 11/11/95 -0800, Simon Spero <ses@tipper.oit.unc.edu> wrote:
>As for real time-release - how about just using conventional encryption,
>and require it to be brute-forced? ..... Moore's law .....
Moore's law is really unpredictable - you can't be sure if the rate of
increase will go up or down, which could affect brute-force time by a
factor of 1000 pretty easily, especially if it suddenly becomes convenient
to do something your crypto-algorithm happens to use.
Also, there's a cost problem - a large brute-force project which requires
N years to crack either needs to be ferociously expensive, or else it's
easy for somebody to put a bunch of machines together to crack it faster.
For anything that requires that level of paranoia, Moore's law probably makes
the timing too unpredictable.
It probably makes a lot more sense to just do contractual solutions,
with secret-sharing protocols to minimize the effects of cheating,
bankrupt service providers, and accidents.
#--
# Thanks; Bill
# Bill Stewart, Freelance Information Architect, stewarts@ix.netcom.com
# Phone +1-510-247-0663 Pager/Voicemail 1-408-787-1281
