Slashdot | 3D Microfluid Computers Used To Solve NP Problems

[Bill Stewart]

At 09:58 PM 03/24/2001 -0800, Ray Dillinger wrote:
>On Sat, 24 Mar 2001, Jim Choate wrote:
> >http://slashdot.org/articles/01/03/24/1840252.shtml
>
>Cryptographically interesting.  It looks like starting now,
>the highest-end threat facing a cryptosystem involves liters
>of fluid performing molecular computation.

Cryptographically *un*interesting.
         > [Good discussion of key lengths, deleted...]

The microfluid computers were solving small instances of NP problems;
larger volumes would let them solve larger instances, but not
blazingly large, though you could get respectably large by using
an ocean-full of the stuff.  That gets you a few extra bits of solution.
But it's not solving general-purpose computing - it's solving
specialized problems which may be somewhat useful for small
instances of NP-hard problems, so it may be useful for
possibly-NP-hard problems like factoring and discrete logs.
Adding a few bits of key length to those problems is less annoying
than adding them to most symmetric-key algorithms, where
there's more likely to be structure to the key length
(e.g. adding a few key bits to RC4 is no problem,
as long as you're below 255, but adding them to DES doesn't work.)

Judging from the abstract, it's still cool stuff,
even if it's not all that practical.