Friam

Why "true" random?

Classic

List

Threaded

11 messages Options

Peter Lissaman

Why "true" random?

Why is it important (except intellectually) to have "true" randomness??? I very well remember the early, good old, bad old, days of Aerospace, in the 50's, when we were really doing practical earthshattering things -- like going to the moon -- sans computers!! The RAND corporation, for whom I consulted, published a typed book (size of a Manhattan telephone directory) of "random" numbers for engineering application. Much entertainment was occasioned when, about three months later, they distributed a list of "typos" to their original list of random numbers. Today I use homemade random numbers alla time for real problems, specifically the actual response of real flight vehicles in real atmospheric turbulence. Flight tests support analysis, in the sense that what we predict is not obviously incorrect. We have never found it necessary to utilize any more "perfectly random" "random" sequences!

Peter Lissaman, Da Vinci Ventures

Expertise is not knowing everything, but knowing what to look for.

1454 Miracerros Loop South, Santa Fe, New Mexico 87505
TEL: (505) 983-7728 FAX: (505) 983-1694
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20070721/e01d235d/attachment.html

Russell Standish

Why "true" random?

Cryptographic applications require true randomness. If your cipher
used on a pseudo-random number generator, then a cracker discovering
your algorithm and key has broken your code.

I also have a hunch that genuine randomness is needed for open-ended
evolutionary systems. Here, the evol algorithm is in the position of
the code cracker, and once the code is cracked, the evol algorithm
stops. I had a workshop paper on this in 2004, which has some problems
with it. The concept is controversial, to say the least.

Cheers

On Sat, Jul 21, 2007 at 10:24:42AM -0600, Peter Lissaman wrote:

> Why is it important (except intellectually) to have "true" randomness??? I very well remember the early, good old, bad old, days of Aerospace, in the 50's, when we were really doing practical earthshattering things -- like going to the moon -- sans computers!! The RAND corporation, for whom I consulted, published a typed book (size of a Manhattan telephone directory) of "random" numbers for engineering application. Much entertainment was occasioned when, about three months later, they distributed a list of "typos" to their original list of random numbers. Today I use homemade random numbers alla time for real problems, specifically the actual response of real flight vehicles in real atmospheric turbulence. Flight tests support analysis, in the sense that what we predict is not obviously incorrect. We have never found it necessary to utilize any more "perfectly random" "random" sequences!
>
>
> Peter Lissaman, Da Vinci Ventures
>
> Expertise is not knowing everything, but knowing what to look for.
>
> 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> TEL: (505) 983-7728 FAX: (505) 983-1694
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org

Robert Holmes

Why "true" random?

In reply to this post by Peter Lissaman

Cryptography. The required robustness of a random generator is highly
sensitive to the intended application;

- Generating a "thought for the day" for your blog? Required
randomness = low.
- Response testing a missile system? Required randomness = medium
- Stealing above test results, encrypting them and transmitting them
to Al Quaeda in a form that you hope the NSA won't understand? Required
randomness = high

Robert

On 7/21/07, Peter Lissaman <plissaman at earthlink.net> wrote:

>
> Why is it important (except intellectually) to have "true" randomness???
> I very well remember the early, good old, bad old, days of Aerospace, in the
> 50's, when we were really doing practical earthshattering things -- like
> going to the moon -- sans computers!! The RAND corporation, for whom I
> consulted, published a typed book (size of a Manhattan telephone directory)
> of "random" numbers for engineering application. Much entertainment was
> occasioned when, about three months later, they distributed a list of
> "typos" to their original list of random numbers. Today I use homemade
> random numbers alla time for real problems, specifically the actual response
> of real flight vehicles in real atmospheric turbulence. Flight tests
> support analysis, in the sense that what we predict is not obviously
> incorrect. We have never found it necessary to utilize any more "perfectly
> random" "random" sequences!
>
>
> Peter Lissaman, Da Vinci Ventures
>
> Expertise is not knowing everything, but knowing what to look for.
>
> 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> TEL: (505) 983-7728 FAX: (505) 983-1694
>
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20070721/bd58347a/attachment.html

Prof David West

Why "true" random?

" cryptography ... missile system ... encrypting ... transmitting ...
Al Quaeda ... NSA" sequence occurring twice within 7 hours in the same
mail-list. Somewhere in VA a computer just burped. Expect the black
helicopters within 24 hours. :)

davew

On Sat, 21 Jul 2007 13:34:44 -0600, "Robert Holmes"
<robert at holmesacosta.com> said:

> Cryptography. The required robustness of a random generator is highly
> sensitive to the intended application;
>
> - Generating a "thought for the day" for your blog? Required
> randomness = low.
> - Response testing a missile system? Required randomness = medium
> - Stealing above test results, encrypting them and transmitting them
> to Al Quaeda in a form that you hope the NSA won't understand?
> Required
> randomness = high
>
> Robert
>
> On 7/21/07, Peter Lissaman <plissaman at earthlink.net> wrote:
> >
> > Why is it important (except intellectually) to have "true" randomness???
> > I very well remember the early, good old, bad old, days of Aerospace, in the
> > 50's, when we were really doing practical earthshattering things -- like
> > going to the moon -- sans computers!! The RAND corporation, for whom I
> > consulted, published a typed book (size of a Manhattan telephone directory)
> > of "random" numbers for engineering application. Much entertainment was
> > occasioned when, about three months later, they distributed a list of
> > "typos" to their original list of random numbers. Today I use homemade
> > random numbers alla time for real problems, specifically the actual response
> > of real flight vehicles in real atmospheric turbulence. Flight tests
> > support analysis, in the sense that what we predict is not obviously
> > incorrect. We have never found it necessary to utilize any more "perfectly
> > random" "random" sequences!
> >
> >
> > Peter Lissaman, Da Vinci Ventures
> >
> > Expertise is not knowing everything, but knowing what to look for.
> >
> > 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> > TEL: (505) 983-7728 FAX: (505) 983-1694
> >
> >
> >
> > ============================================================
> > FRIAM Applied Complexity Group listserv
> > Meets Fridays 9a-11:30 at cafe at St. John's College
> > lectures, archives, unsubscribe, maps at http://www.friam.org
> >

Phil Henshaw-2

Why "true" random?

In reply to this post by Russell Standish

Or what about 'decynchronization', rather than random noise..to erase
inconvenient pattern? Probably has nothing to do with cryptography,
though, I suppose, as I expect that the sort of lab experiment thing the
people at the SASO conference were talking about has no mathematical
representation as yet, just ways of producing them. At least that's
another property that efficiently hides pattern. It came up that some
of the work on syncronization, that doing the opposite had valuable
proprerties in preventing congestion and surges when used to produce
desynchronized flows. Interesting work though!

Phil Henshaw ????.?? ? `?.????
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
680 Ft. Washington Ave
NY NY 10040
tel: 212-795-4844
e-mail: pfh at synapse9.com
explorations: www.synapse9.com

> -----Original Message-----
> From: friam-bounces at redfish.com
> [mailto:friam-bounces at redfish.com] On Behalf Of Russell Standish
> Sent: Saturday, July 21, 2007 12:28 AM
> To: plissaman at earthlink.net; The Friday Morning Applied
> Complexity Coffee Group
> Subject: Re: [FRIAM] Why "true" random?
>
>
> Cryptographic applications require true randomness. If your
> cipher used on a pseudo-random number generator, then a
> cracker discovering your algorithm and key has broken your code.
>
> I also have a hunch that genuine randomness is needed for
> open-ended evolutionary systems. Here, the evol algorithm is
> in the position of the code cracker, and once the code is
> cracked, the evol algorithm stops. I had a workshop paper on
> this in 2004, which has some problems with it. The concept is
> controversial, to say the least.
>
> Cheers
>
> On Sat, Jul 21, 2007 at 10:24:42AM -0600, Peter Lissaman wrote:
> > Why is it important (except intellectually) to have "true"
> > randomness??? I very well remember the early, good old,
> bad old, days
> > of Aerospace, in the 50's, when we were really doing practical
> > earthshattering things -- like going to the moon -- sans
> computers!!
> > The RAND corporation, for whom I consulted, published a typed book
> > (size of a Manhattan telephone directory) of "random" numbers for
> > engineering application. Much entertainment was occasioned when,
> > about three months later, they distributed a list of
> "typos" to their
> > original list of random numbers. Today I use homemade
> random numbers
> > alla time for real problems, specifically the actual
> response of real
> > flight vehicles in real atmospheric turbulence. Flight
> tests support
> > analysis, in the sense that what we predict is not obviously
> > incorrect. We have never found it necessary to utilize any more
> > "perfectly random" "random" sequences!
> >
> >
> > Peter Lissaman, Da Vinci Ventures
> >
> > Expertise is not knowing everything, but knowing what to look for.
> >
> > 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> > TEL: (505) 983-7728 FAX: (505) 983-1694
> > ============================================================
> > FRIAM Applied Complexity Group listserv
> > Meets Fridays 9a-11:30 at cafe at St. John's College lectures,
> > archives, unsubscribe, maps at http://www.friam.org
>
> --
>
> --------------------------------------------------------------
> --------------
> A/Prof Russell Standish Phone 0425 253119 (mobile)
> Mathematics
> UNSW SYDNEY 2052 hpcoder at hpcoders.com.au
> Australia http://www.hpcoders.com.au
> --------------------------------------------------------------
> --------------
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org
>
>

Douglas Roberts-2

Why "true" random?

In reply to this post by Russell Standish

Simulations of stochastic processes also require good RN generators,
especially for simulations of large systems with (I hate to use this word)
emergent behavioral properties. A bad RN generator will introduce emergent
behavior that will be "flavored" by a bad random sequences.

--
Doug Roberts, RTI International
droberts at rti.org
doug at parrot-farm.net
505-455-7333 - Office
505-670-8195 - Cell

On 7/20/07, Russell Standish <r.standish at unsw.edu.au> wrote:

>
> Cryptographic applications require true randomness. If your cipher
> used on a pseudo-random number generator, then a cracker discovering
> your algorithm and key has broken your code.
>
> I also have a hunch that genuine randomness is needed for open-ended
> evolutionary systems. Here, the evol algorithm is in the position of
> the code cracker, and once the code is cracked, the evol algorithm
> stops. I had a workshop paper on this in 2004, which has some problems
> with it. The concept is controversial, to say the least.
>
> Cheers
>
> On Sat, Jul 21, 2007 at 10:24:42AM -0600, Peter Lissaman wrote:
> > Why is it important (except intellectually) to have "true"
> randomness??? I very well remember the early, good old, bad old, days of
> Aerospace, in the 50's, when we were really doing practical earthshattering
> things -- like going to the moon -- sans computers!! The RAND corporation,
> for whom I consulted, published a typed book (size of a Manhattan telephone
> directory) of "random" numbers for engineering application. Much
> entertainment was occasioned when, about three months later, they
> distributed a list of "typos" to their original list of random
> numbers. Today I use homemade random numbers alla time for real problems,
> specifically the actual response of real flight vehicles in real atmospheric
> turbulence. Flight tests support analysis, in the sense that what we
> predict is not obviously incorrect. We have never found it necessary to
> utilize any more "perfectly random" "random" sequences!
> >
> >
> > Peter Lissaman, Da Vinci Ventures
> >
> > Expertise is not knowing everything, but knowing what to look for.
> >
> > 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> > TEL: (505) 983-7728 FAX: (505) 983-1694
> > ============================================================
> > FRIAM Applied Complexity Group listserv
> > Meets Fridays 9a-11:30 at cafe at St. John's College
> > lectures, archives, unsubscribe, maps at http://www.friam.org
>
> --
>
>
> ----------------------------------------------------------------------------
> A/Prof Russell Standish Phone 0425 253119 (mobile)
> Mathematics
> UNSW SYDNEY 2052 hpcoder at hpcoders.com.au
> Australia http://www.hpcoders.com.au
>
> ----------------------------------------------------------------------------
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20070721/bb754afb/attachment.html

Phil Henshaw-2

Why "true" random?

Not sure really what the inputs always used, but I think these Self-org
& Self-adapt algorithms the SASO engineers were playing with didn't
always use random generators to produce the systemic effects they were
getting. Obviously all input effects all output in some sort of way,
but it was the outcomes that would come from the whole gamete of
unspecified inputs that seemed to be the 'phase space profile' they were
most interested in.

Many of the papers were on how the inputs could seriously 'misbehave'
and still not screw up the control schemes, often discussed in terms of
'malicious agent' concepts, of which the real net has plenty real
examples! I also found them very receptive to considering not only
what a malicious person would think of doing to defeat someone else's
operating plan, but also the 'malicious creativity' of natural system
emergence as a focus of design contingencies.

Phil Henshaw ????.?? ? `?.????
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
680 Ft. Washington Ave
NY NY 10040
tel: 212-795-4844
e-mail: pfh at synapse9.com
explorations: www.synapse9.com <http://www.synapse9.com/>

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On
Behalf Of Douglas Roberts
Sent: Sunday, July 22, 2007 1:19 AM
To: The Friday Morning Applied Complexity Coffee Group;
plissaman at earthlink.net
Subject: Re: [FRIAM] Why "true" random?

Simulations of stochastic processes also require good RN generators,
especially for simulations of large systems with (I hate to use this
word) emergent behavioral properties. A bad RN generator will introduce
emergent behavior that will be "flavored" by a bad random sequences.

--
Doug Roberts, RTI International
droberts at rti.org
doug at parrot-farm.net
505-455-7333 - Office
505-670-8195 - Cell

On 7/20/07, Russell Standish <r.standish at unsw.edu.au> wrote:

Cryptographic applications require true randomness. If your cipher
used on a pseudo-random number generator, then a cracker discovering
your algorithm and key has broken your code.

I also have a hunch that genuine randomness is needed for open-ended
evolutionary systems. Here, the evol algorithm is in the position of
the code cracker, and once the code is cracked, the evol algorithm
stops. I had a workshop paper on this in 2004, which has some problems
with it. The concept is controversial, to say the least.

Cheers

On Sat, Jul 21, 2007 at 10:24:42AM -0600, Peter Lissaman wrote:
> Why is it important (except intellectually) to have "true"
randomness??? I very well remember the early, good old, bad old, days
of Aerospace, in the 50's, when we were really doing practical
earthshattering things -- like going to the moon -- sans computers!!
The RAND corporation, for whom I consulted, published a typed book (size
of a Manhattan telephone directory) of "random" numbers for engineering
application. Much entertainment was occasioned when, about three months
later, they distributed a list of "typos" to their original list of
random numbers. Today I use homemade random numbers alla time for real
problems, specifically the actual response of real flight vehicles in
real atmospheric turbulence. Flight tests support analysis, in the
sense that what we predict is not obviously incorrect. We have never
found it necessary to utilize any more "perfectly random" "random"
sequences!

>
>
> Peter Lissaman, Da Vinci Ventures
>
> Expertise is not knowing everything, but knowing what to look for.
>
> 1454 Miracerros Loop South, Santa Fe, New Mexico 87505
> TEL: (505) 983-7728 FAX: (505) 983-1694
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> lectures, archives, unsubscribe, maps at http://www.friam.org

--

------------------------------------------------------------------------
----
A/Prof Russell Standish Phone 0425 253119 (mobile)
Mathematics
UNSW SYDNEY 2052 hpcoder at hpcoders.com.au
Australia http://www.hpcoders.com.au
<http://www.hpcoders.com.au>
------------------------------------------------------------------------
----

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20070722/bef1457b/attachment.html

Roger Frye-3

Why "true" random?

In reply to this post by Douglas Roberts-2

I would argue the opposite. While I agree with Doug that you need good
RNGs (though not necessarily true RNGs) in order to avoid bias, the
problem with good pseudo- or true- RNGs is that they have order N^2
convergence for Monte Carlo simulations. Quasi-random number generators
on the other hand (such as multiples of an irrational square root, or a
Peano tiling) converge in order N. If you can trust the results, faster
conergence lets you simulate more.
-Roger

On Sat, 21 Jul 2007 23:18:36 -0600, Douglas Roberts <doug at parrot-farm.net>
wrote:

> Simulations of stochastic processes also require good RN generators,
> especially for simulations of large systems with (I hate to use this
> word)
> emergent behavioral properties. A bad RN generator will introduce
> emergent
> behavior that will be "flavored" by a bad random sequences.
>
>

Robert Howard-2-3

Why "true" random?

How about deleting confidential data from hard disks!

The solution today is to overwrite it many times with random data.

But modern mathematics and technology makes it possible to recover the much
of the original text given the original random sequence used to delete the
data. Given a long sequence of deleted white space (or zeros on the disk),
then it becomes possible to recover the original pseudo-random sequence (for
example, one based on linear
<http://en.wikipedia.org/wiki/Linear_congruence_theorem> congruence) - even
if a many passes are performed.

With a true random number generator, only one pass is needed.

I'm sure hardware random number generators based on quantum effects, which
have been around for decades, would be used instead of hitting a web site,
which compromises the who shebang.

Robert Howard

Phoenix, Arizona

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf
Of Roger Frye
Sent: Sunday, July 22, 2007 6:40 AM
To: The Friday Morning Applied Complexity Coffee Group
Subject: Re: [FRIAM] Why "true" random?

I would argue the opposite. While I agree with Doug that you need good

RNGs (though not necessarily true RNGs) in order to avoid bias, the

problem with good pseudo- or true- RNGs is that they have order N^2

convergence for Monte Carlo simulations. Quasi-random number generators

on the other hand (such as multiples of an irrational square root, or a

Peano tiling) converge in order N. If you can trust the results, faster

conergence lets you simulate more.

-Roger

On Sat, 21 Jul 2007 23:18:36 -0600, Douglas Roberts <doug at parrot-farm.net>

wrote:

> Simulations of stochastic processes also require good RN generators,

> especially for simulations of large systems with (I hate to use this

> word)

> emergent behavioral properties. A bad RN generator will introduce

> emergent

> behavior that will be "flavored" by a bad random sequences.

>

>

============================================================

FRIAM Applied Complexity Group listserv

Meets Fridays 9a-11:30 at cafe at St. John's College

lectures, archives, unsubscribe, maps at http://www.friam.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://redfish.com/pipermail/friam_redfish.com/attachments/20070722/00236589/attachment.html

Russell Standish

Why "true" random?

On Sun, Jul 22, 2007 at 01:52:35PM -0700, Robert Howard wrote:

> How about deleting confidential data from hard disks!
>
> The solution today is to overwrite it many times with random data.
>
> But modern mathematics and technology makes it possible to recover the much
> of the original text given the original random sequence used to delete the
> data. Given a long sequence of deleted white space (or zeros on the disk),
> then it becomes possible to recover the original pseudo-random sequence (for
> example, one based on linear
> <http://en.wikipedia.org/wiki/Linear_congruence_theorem> congruence) - even
> if a many passes are performed.
>
> With a true random number generator, only one pass is needed.
>
> I'm sure hardware random number generators based on quantum effects, which
> have been around for decades, would be used instead of hitting a web site,
> which compromises the who shebang.
>
>
>
> Robert Howard
>
> Phoenix, Arizona
>
>

You don't even need to do that. Entropic sources are available from
considering timings on a system undergoing interrupts from external
sources (eg mouse or keyboard activity). The Linux kernel performs
this analysis and provides a conveniently encapsulated device called
/dev/random. I used used precisely this technique to implement a disk
erasing program a couple of years ago - and offered the possibility
to do it multiple times for the absolutely paranoid.

Note that /dev/random has rather unpredictable performance - you are
advised to shake you mouse, or something like that when generating a
seed for ssh for instance. To improve its performance, you use the
output of /dev/random to fill a table, which is continually
overwritten as new random bits become available, Then you use a
conventional pseudo RNG to index into the table, so the resulting
bitstream has small chunks of "correlated" numbers, but is by and
large unpredictable.

The state of the art for doing this is a library called Havege. Look
it up if you're interested.

Cheers

--

----------------------------------------------------------------------------
A/Prof Russell Standish Phone 0425 253119 (mobile)
Mathematics
UNSW SYDNEY 2052 hpcoder at hpcoders.com.au
Australia http://www.hpcoders.com.au
----------------------------------------------------------------------------

James Steiner

Why "true" random?

In reply to this post by Robert Howard-2-3

Robert,

It's my understanding that there has been no documented case of data
recovered from a hard disk that has been erased by completely
overwriting the contents 3 or more times with your choice of 0s, 1s,
alternating bits, random bits, or whatever, outside of a lab
environment using magnetic electron microscopy. Is that no longer
true? I had thought that one didn't need a particularly good RNG for
it, since anything will do. Or, is that just what the NSA *wants* me
to think?

~~James
_____________________
http://www.turtlezero.com

On 7/22/07, Robert Howard <rob at symmetricobjects.com> wrote:

> How about deleting confidential data from hard disks!
>
> The solution today is to overwrite it many times with random data.
>
> But modern mathematics and technology makes it possible to recover the much
> of the original text given the original random sequence used to delete the
> data. Given a long sequence of deleted white space (or zeros on the disk),
> then it becomes possible to recover the original pseudo-random sequence (for
> example, one based on linear congruence) ? even if a many passes are
> performed.
>
> With a true random number generator, only one pass is needed.
>
> I'm sure hardware random number generators based on quantum effects, which
> have been around for decades, would be used instead of hitting a web site,
> which compromises the who shebang.
>
>
>
> Robert Howard
>
> Phoenix, Arizona