At 03:26 PM 9/24/96 +0200, Gary Howland <gary@systemics.com> wrote:

I fail to see *any* (non educational) use for these DNA "computers", let alone a cryptographic use - sure, they may be massively parallel, but what's the big deal? I can now perform a calculation a million times faster than I could yesterday? (something I personally doubt, but will agree to for sake of the argument).

One mole of a substance contains ~6x10**23 molecules, and weighs one gram per atomic-weight of the molecule. A DNA "computer" might weigh a kilo or two for one mole of computer virus. It may not be blazingly fast, especially if you've got to synthesize lots of different molecules to make it up, plus extract the result from the data soup, but 10**23 is a _big_ number. It's probably not very useful for cryptographic applications, but Adleman was using it to solve Travelling Salesman problems, which are NP-hard, and if you do have a crypto problem that maps well into TSP, a hot-tub full of interesting solutions might be an interesting solution, especially if you've got a huge underground lab and a National Institutes of Health nearby in case you have any bugs or memory leaks in your program..... Unlike quantum computing, it doesn't change the exponentiality of the problems it's solving, it just multiplies the computing capability by a very big constant, and it does parallelize cleanly. The engineering is tricky, but I find it more believable than practical high-precision quantum computers. And if the DNA computer doesn't work, you can always recycle your lab to synthesize large quantities of recreational pharamceuticals and use the money from them to bribe the person who knows the key. Avogadro # Thanks; Bill # Bill Stewart, +1-415-442-2215 stewarts@ix.netcom.com # <A HREF="http://idiom.com/~wcs"> # You can get PGP software outside the US at ftp.ox.ac.uk/pub/crypto