On Tue, 23 Apr 2002 [EMAIL PROTECTED] wrote:
> -- > Jim Choate wrote: > > > > If you can't develop a RNG in software (ie you'd be in a > > > > state of sin), what makes you think you can do it using > > > > -only- digital gates in hardware? You can't. > > James A. Donald: > > > Classic Choatian physics. > > > > > > Of course you can. > > Jim Choate: > > Not if you use -only- digital gates and derived functions (eg > > flip flop or a shift register) > > One can build a true random generator using a two resistors, a > capacitor, three unbuffered inverters and a shift register. A > second shift register and an XOR gate will serve to quite > adequately whiten the output. Yeah, the shit for brains will probably tell you that resistors and capacitors aren't digital gates. In which case you can just use a bunch of flip-flop that feed back to themselves at different rates, so you get different streams of 1's and 0's, and xor'em all together. Inefficient as compared to using analog components, but it would work if the timing between them is different enough and you only get data from them sporadically. Lots of whitening to do afterwards of course...
