RE: [boost] Random numbers

From: Reid Sweatman (drunkardswalk_at_[hidden])
Date: 2003-09-11 07:35:00

Okay, then you *are* doing something like Yarrow.  Have you read that paper?
I always liked the idea myself, and I'm sure someone here will jump in to
remind me why no one else did <g>.  On a side note, I had an interesting
conversation with one of AMD's design people a couple of years back where I
suggested that they implement a white noise generator with an opcode to
sample it right on the CPU silicon.  I don't think it had occurred to them
before that, and, of course, they said it would go on the wish list, but I'm
guessing it got thirty-filed right away.  Still, it would be a nice thing to
have, and arguably more useful than lots of the things they waste opcode
space on.

Anyway, I seem to be the only one responding, at least on the list.  I like
the idea, especially if it can be faired into the existing random module
cleanly.  One last question:  since so many of the good sources of entropy
on a machine are platform-dependent, how are you addressing cross-platform
portability?  And with that, I'll bow out of the thread and make room for
others to get a word in edgewise.

Reid Sweatman

> -----Original Message-----
> From: boost-bounces_at_[hidden]
> [mailto:boost-bounces_at_[hidden]]On Behalf Of Baris KAYGISIZER
> Sent: Thursday, September 11, 2003 1:05 AM
> To: Boost mailing list
> Subject: Re: [boost] Random numbers
>
>
> I think, I don't explain. sorry. Absolutely, my generator is using
> outside/external source too, if i want to procreate real random
> numbers. My
> outside/external device is collecting source as voice/sound/noise,
> picture/photo, movie and etc for generator. Generator/Creator is a little
> computer program on a PC. Generator confuses sources comes from collector
> with computer program. And then creates random numbers. Thats all shortly.
>
> Baris.
>
>
> ----- Original Message -----
> From: "Reid Sweatman" <drunkardswalk_at_[hidden]>
> To: "Boost mailing list" <boost_at_[hidden]>
> Sent: Wednesday, September 10, 2003 8:29 AM
> Subject: RE: [boost] Random numbers
>
>
> > Since it's already top-posted, I'll keep it that way, or it'll
> really get
> > confusing.  I think I'm a little more clear on what you're
> talking about,
> > but still not sure.  Okay, no external device, just software.  But not a
> > pseudo-random number generator (PRNG), i.e., one that, given the same
> > starting state, always produces the same sequence of numbers, right?  To
> do
> > that, you *have* to have some source of entropy, and you've just said,
> it's
> > not hardware.  That seems to leave something like Schneier's Yarrow,
> right?
> > Or am I still missing something?  There was a thread on the topic some
> > little while ago that should turn up in the archives.  As I recall, I
> > proposed including Yarrow, and others raised some pretty good points in
> > rebuttal, so it never happened.  Is this the kind of thing
> you're talking
> > about?  Using various mensurable things on the computer, such as timing
> > variations between keystrokes, variations in network response times, the
> > contents of various machine registers and memory locations at particular
> > times (many ways to vary when they're read), to feed an "entropy pool,"
> > which is churned continuously and tapped as needed?  This is the basic
> idea
> > behind Yarrow, the original paper on which should still be available at
> > www.counterpane.com.  Or am I misunderstanding you still?  I'm
> confused in
> > that you say you're "not using noise when creating."  There
> *must* be some
> > source of entropy, of randomness, since, as you noted, any
> purely software
> > algorithm is completely deterministic.  (Incidentally,
> reproducibility of
> a
> > "random" stream is a plus in some kinds of application).
> >
> > Reid
> >
> > > -----Original Message-----
> > > From: boost-bounces_at_[hidden]
> > > [mailto:boost-bounces_at_[hidden]]On Behalf Of Baris KAYGISIZER
> > > Sent: Tuesday, September 09, 2003 4:26 AM
> > > To: Boost mailing list
> > > Subject: Re: [boost] Random numbers
> > >
> > >
> > > Hi Reid,
> > >
> > > Yes, i am using Outlook Express. i defined plain text this e-mail. I
> hope,
> > > no problem. That i mean; Some computer programs generate
> random numbers
> > > without using outside device/machine/or whatever. These random numbers
> are
> > > not real, not casual. Because, only computer programs generate same
> random
> > > numbers always. True random number generators are using noise
> amplifier,
> > > thermal noise, etc. that i read lots of article. This is the
> > > similer that i
> > > made my generator. I am not using noise when creating.
> > >
> > > kindest
> > >
> > > Baris.
> > >
> > >
> > > ----- Original Message -----
> > > From: Reid Sweatman
> > > To: Boost mailing list
> > > Sent: Tuesday, September 09, 2003 3:13 AM
> > > Subject: RE: [boost] Random numbers
> > >
> > >
> > > First, sorry for the top-post; Outlook likes them, and it's hard to
> defeat
> > > its HTML formatting.  On to the meat.  What do you mean when you
> > > say "Random
> > > numbers sources are not computer program"?  That you're using machine
> > > properties as an entropy source?  Or that you have an attached device
> that
> > > supplies entropy, say like something that amplifies
> transistor junction
> > > noise?  It's hard to evaluate your proposal when you're not sure what
> it's
> > > proposing.
> > >
> > > Reid Sweatman
> > > -----Original Message-----
> > > From: boost-bounces_at_[hidden]
> > > [mailto:boost-bounces_at_[hidden]]On
> > > Behalf Of Baris KAYGISIZER
> > > Sent: Monday, September 08, 2003 8:54 AM
> > > To: boost_at_[hidden]
> > > Subject: [boost] Random numbers
> > >
> > >
> > > Hi all,
> > >
> > >
> > > i have a question. i am working on a random numbers in last 2
> > > weeks. Who can
> > > use the random numbers or who wants? I can create real random
> > > numbers. Real
> > > random numbers mean: Random numbers sources are not computer program.
> Who
> > > wants to use these?
> > >
> > > Thx.
> > >
> > >
> > > Baris KAYGISIZER
> > > baris.kaygisizer_at_[hidden]
> > >
> > >
> > > _______________________________________________
> > > Unsubscribe & other changes:
> > > http://lists.boost.org/mailman/listinfo.cgi/boost
> > >
> >
> >
> > _______________________________________________
> > Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
> >
>
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost