Re: Random Number Generator

I know it's a dead thread, but was just too busy at the time. Still, it's such 
an interesting area, I had to add my $.02 even at this late date.

First, any thought about _truly_ random numbers must begin from somewhere like 
this: Given that the number line is infinite, what is the probability that any 
_truly_ random number will be small enough to express within the volume of the 
known universe?

And then, most of the confusion comes from mixed definitions and usages. At 
some points in this thread, the talk was about algorithms (i.e., programs) to 
generate 'random' 'numbers'. At other points, it was about predicting outcomes 
of real-world events (e.g., a coin-toss). The two are not interchangeable. In 
short, a program will never generate truly random numbers even from a large set 
of numbers unless there is a random seed. For all practical purposes, the only 
way to obtain a random seed is through monitoring for a random real-world event.

As far as generating PIN numbers, what difference does it make? Sequential 
numbers might be almost as useful. Without knowing significant details such as 
the volume of numbers generated within a known time period and other elements, 
who could guess the generated number anyway? Since sequentially generating them 
will provide a nice, neat distribution over the set, what more do you need? 
Maybe add the digits position from QSECOND or something just to keep it 
interesting; it just probably doesn't matter.

Tom Liotta


On Wed, 08 August 2001, "alan shore" wrote:

>
> Well Leif, it looks like the only people interested in this discussion is you 
>and me. So if you agree, we will each have one more response, before we start 
>to hear the boo's and hisses from the audience.
> In reading your last response, I believe you contradicted yourself. The 
>"randomness" of a number is dependent upon the number of numbers available 
>from your machine. In your example, a coin, the fact that you have 2 choices 
>does not give a large sample, BUT, is the result of the first flip known 
>beforehand. The answer is NO. Can you guess it? The answer is a 50-50 chance. 
>The greater the number of choices, the harder it is to guess that initial 
>result, but are both these examples random? The answer is most definitely yes.

--
Tom Liotta
The PowerTech Group, Inc.
19426 68th Avenue South
Kent, WA 98032
Phone  253-872-7788
Fax  253-872-7904
http://www.400Security.com


___________________________________________________
The ALL NEW CS2000 from CompuServe
 Better!  Faster! More Powerful!
 250 FREE hours! Sign-on Now!
 http://www.compuserve.com/trycsrv/cs2000/webmail/