| IDL random number generator [message #35041] |
Fri, 09 May 2003 04:47  |
krijger
Messages: 5
Registered: May 2003
|
Junior Member |
|
|
Hi,
I know that randomn is pseudo-random, how many numbers can you
generate before the non-randomness kicks in?
Thijs Krijger
|
|
|
|
|
|
|
|
| Re: IDL random number generator [message #35129 is a reply to message #35041] |
Tue, 20 May 2003 19:54  |
tandp
Messages: 8
Registered: November 2000
|
Junior Member |
|
|
> influenced by the temperature. I'll admit, it's a subtle difference, but
> if you were doing serious scientific work using those random numbers,
> it's one you should worry about when designing the program.
For scientific work a pseudorandom number generator should be sufficient
as long as it is well designed, i.e. has cycle longer than the number of
values it will be rquired to generate and generates data that has a
prescribed mean (usually zero). It should conform to the probablility
distribution it is designed for. A random number generator of values from
a uniform distribution should be verifiable as generating data with a zro
mean and have a frequency spectrum indistiguishable from white noise.
|
|
|
|
| Re: IDL random number generator [message #35143 is a reply to message #35041] |
Tue, 20 May 2003 07:30 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
Big Bird wrote:
>
> David Fanning <david@dfanning.com> wrote in message news:<MPG.192c7f24abbae652989b96@news.frii.com>...
>> Folks,
>>
>> You guys might want to check out this Quantum Random
>> Number Generator. This one takes a LONG time to repeat! :-)
>>
>> http://www.idquantique.com/qrng.html
>>
>> Cheers,
>>
>> David
>
> I would mistrust this for the following reasons.
>
> The website states:
>
>> Being deterministic devices, computers are not capable of producing random
>> number generators.
>
> That statement is false: any odd soundcard has a noise-generator
> (essentially a glorified resistor with a big amplifier attached to it)
> that is capable of producing perfectly fine thermal random noise. I've
> used that for creating random numbers since back in the days of the
> Atari-800 (OK, so I'm dating myself here).
It's an exagerration, rather than being false. Their real point was that
the standard random number generators are not truly random, and can't be
truly random as long as they are entirely based in software.
> Even if a piece of external hardware was desirable for this process it
> could be cobbled together for $5 from a couple resistors, an opAmp for
> amplification and an RC high-pass filter (to get rid of slow drifting
> in the opAmp). Potentially another opAmp driven open-loop as a
> Schmitt-trigger. Gives you clock-free(!) bit-noise.
>
> Anybody who tries to sell a quantum-ANYTHING to make noise(!!) is
> direputable from the word 'go' in my eyes.
Thermal noise is not as fundamentally random as the quantum fluctuations
that determine radioactive decay. Among other things, your circuit will
pick up and therefore be influenced by EM radiation, which is not
random, and since it's thermal noise you're looking at, it will be
influenced by the temperature. I'll admit, it's a subtle difference, but
if you were doing serious scientific work using those random numbers,
it's one you should worry about when designing the program.
|
|
|
|
| Re: IDL random number generator [message #35146 is a reply to message #35041] |
Mon, 19 May 2003 21:18 |
rmoss4
Messages: 21
Registered: October 2002
|
Junior Member |
|
|
Big Bird wrote:
> David Fanning <david@dfanning.com> wrote in message news:<MPG.192c7f24abbae652989b96@news.frii.com>...
>
>> Folks,
>>
>> You guys might want to check out this Quantum Random
>> Number Generator. This one takes a LONG time to repeat! :-)
>>
>> http://www.idquantique.com/qrng.html
>>
>> Cheers,
>>
>> David
>
>
> I would mistrust this for the following reasons.
>
> The website states:
>
>
>> Being deterministic devices, computers are not capable of producing random
>> number generators.
>
>
> That statement is false: any odd soundcard has a noise-generator
> (essentially a glorified resistor with a big amplifier attached to it)
> that is capable of producing perfectly fine thermal random noise. I've
> used that for creating random numbers since back in the days of the
> Atari-800 (OK, so I'm dating myself here).
>
> Even if a piece of external hardware was desirable for this process it
> could be cobbled together for $5 from a couple resistors, an opAmp for
> amplification and an RC high-pass filter (to get rid of slow drifting
> in the opAmp). Potentially another opAmp driven open-loop as a
> Schmitt-trigger. Gives you clock-free(!) bit-noise.
>
> Anybody who tries to sell a quantum-ANYTHING to make noise(!!) is
> direputable from the word 'go' in my eyes.
Thermal random noise is itself quantum in nature. Though not stated very
precisely, I think the web site is referring to computer software (as
opposed to hardware) as being deterministic and therefore non-random. I
agree with you though that there are less expensive ways of making noise :)
--
Robert M. Moss, PhD
|
|
|
|
| Re: IDL random number generator [message #35149 is a reply to message #35076] |
Mon, 19 May 2003 16:51 |
condor
Messages: 35
Registered: January 2002
|
Member |
|
|
David Fanning <david@dfanning.com> wrote in message news:<MPG.192c7f24abbae652989b96@news.frii.com>...
> Folks,
>
> You guys might want to check out this Quantum Random
> Number Generator. This one takes a LONG time to repeat! :-)
>
> http://www.idquantique.com/qrng.html
>
> Cheers,
>
> David
I would mistrust this for the following reasons.
The website states:
> Being deterministic devices, computers are not capable of producing random
> number generators.
That statement is false: any odd soundcard has a noise-generator
(essentially a glorified resistor with a big amplifier attached to it)
that is capable of producing perfectly fine thermal random noise. I've
used that for creating random numbers since back in the days of the
Atari-800 (OK, so I'm dating myself here).
Even if a piece of external hardware was desirable for this process it
could be cobbled together for $5 from a couple resistors, an opAmp for
amplification and an RC high-pass filter (to get rid of slow drifting
in the opAmp). Potentially another opAmp driven open-loop as a
Schmitt-trigger. Gives you clock-free(!) bit-noise.
Anybody who tries to sell a quantum-ANYTHING to make noise(!!) is
direputable from the word 'go' in my eyes.
|
|
|
|
| Re: IDL random number generator [message #35216 is a reply to message #35076] |
Thu, 15 May 2003 08:19 |
rmoss4
Messages: 21
Registered: October 2002
|
Junior Member |
|
|
That is pretty cool. However, it would not appear to be very useful if
you need the ability to replicate your results (e.g. in cases where you
would use a known seed value to reproduce a pseudo-random series of
numbers). Nevertheless, it's probably as close as one can get to "true"
randomness. Ain't physics great?
--
Robert M. Moss, PhD
rmoss4@houston.rr.com
281-856-2017
David Fanning wrote:
> Folks,
>
> You guys might want to check out this Quantum Random
> Number Generator. This one takes a LONG time to repeat! :-)
>
> http://www.idquantique.com/qrng.html
>
> Cheers,
>
> David
>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
