From: Michael Rhoades (mrhoades@perceptionfactory.com)
Date: Fri Apr 08 2011 - 15:09:05 EDT
I ascribe to the notion that true randomness is probably not possible.
Yet patterns that are beyond our comprehension, even using the current
tools of computation at our disposal, are abundant. A formula for Pi for
instance... "Degree" of randomness, previously alluded to, is an
oxymoron and might be better termed degree of comprehension. We can only
base the idea of randomness upon our own inability to compute a patterns
in relation to phenomena.
On 4/8/11 2:03 PM, Michael Gogins wrote:
> True randomness is an extremely uncanny thing. It has been proved that
> there is no way to produce true randomness with a computer, and that
> there is no way to use a computer to prove that an infinite sequence
> is truly random. Consequently if Nature contains phenomena that are
> truly random (as for example quantum mechanics predicts), then Nature
> cannot be perfectly simulated by a computer.It can perhaps be
> simulated "as closely as one likes", but that is not quite the same
> thing as "perfectly." (And it has been proved that the probability of
> a random program halting is an uncomputable number.)
>
> Therefore in spite of the case you cite "physical sources of entropy"
> are in fact often used in computing, either as sources of
> cryptographic keys or for random number generation. Steps are in fact
> taken to remove measured biases and so on.
>
> Regards,
> Mike
>
> On Fri, Apr 8, 2011 at 1:38 PM, Richard Dobson
> <richarddobson@blueyonder.co.uk
> <mailto:richarddobson@blueyonder.co.uk>> wrote:
>
> On 08/04/2011 17:23, Michael Rhoades wrote:
>
> Hi Michael.
>
> As soon as I sent that I wished I had said "could be as
> "random"..."
>
> Should bias be included within the realm of constraints? Human
> bias is
> perhaps a result of one's conditioning. Would you feel that a
> program
> written to produce quasi-random numerical output would include
> any form
> of bias?
>
>
> Human bias is largely down to the fact that randomness is not
> generally of much use to humans; the human instinct is very much
> to look for patterns and trends (and doubtless meta-patterns and
> meta-trends), on which predictions that might have a survival or
> other value may be based. Superstition (or "intuition" or "sixth
> sense" etc) is certainly one approach to the problem of making
> choices; but faith in patterns and trends greatly outweighs any
> instinct or habit regarding randomness. Which is why, famously,
> after ten coin tosses have thrown heads, many people (unless they
> have expressly been taught otherwise) will expect the eleventh to
> be more likely to to be heads too, and why most people in choosing
> lottery numbers will likely avoid any suggestion of a pattern
> (even one as simple as two consecutive numbers), and in trying to
> be as "random as possible" will try to distribute their choices
> evenly over the set.
>
> So great is the concern amongst some scientists that a computer
> algorithm cannot reliably produce a "truly" random stream of
> numbers that only recently an experiment was set up to use
> radioactive decay as a source of said "truly random" numbers. All
> was fine an dandy until they discovered, by chance, that the
> stream of numbers from the supposedly constant decay rate had a
> slow periodic pattern to it, that later was found to be correlated
> in some unknown way with solar activity. So a computer algorithm
> may in the end be deemed to be better than some physically based
> source, but the suspicion will probably always remain that,
> actually, there ~is~ some inescapable degree of correlation,
> somewhere.
>
> Richard Dobson
>
>
>
>
> --
> Michael Gogins
> Irreducible Productions
> http://www.michael-gogins.com
> Michael dot Gogins at gmail dot com
This archive was generated by hypermail 2.1.5 : Fri Apr 08 2011 - 22:06:39 EDT