Slashdot Mirror
True Random Number Generator Goes Online
amigoro writes "A 'true' random number generator that relies on the unpredictable quantum process of photon emission has gone online providing academic and scientific community access to true random numbers free of charge."
..BZT, qvq nalbar frr gung gb ertvfgre lbh unir
gb fbyir n zngu ceboyrz yvxr:
qrevingvir bs (5*fva 3k +6pbf(-cv/2))
Avpr!
Urer vf n qverpg yvax gb gur trarengbe, lbh pna
qbjaybnq gur pyvrag sebz urer nf jryy:
uggc://enaqbz.veo.ue/
DEnaq Pbzznaq-yvar Hgvyvgl [i0.2, 2007-07-17]
Abgr 1: Pbzcvyrf haqre Ivfhny Fghqvb naq t++.
Abgr 2: Jvaqbjf rkrphgnoyr vapyhqrq.
Abgr 3: TAH Yvahk rkrphgnoyr vapyhqrq.
Hey! It works!
Why do I keep getting 42?
when I think of random numbers, I can't help but remember the 'fishbowl' that had at SGI (mtn view) where an indycam was photo'ing some lavalamps and creating random seeds based on those images.
ah, SGI....
--
"It is now safe to switch off your computer."
Hasn't random.org done this for a while already? Perhaps they don't have academic backing, but I do believe they use numbers generated by atomic decay.
True random number generators have been around in hardware form for a while based on a number of different processes, not quantum only. But this is being offered to the community at large, who may not have the means to procure or pay for a hardware solution.
This is neat but there have been other quantum random number generators online for years. This one by id Quantique springs to mind... I'm not sure what this new service provides that others don't. If you REALLY want secure random numbers you should buy a QRNG PCI card and make them yourself so you're the only one with a copy.
John Walker (AutoDesk founder) has had a true random number generator available for web access for quite a long time. Looks like his site's currently down, but check out www.fourmilab.ch when it's sorted -- in addition to the random number generator he has a number of other cool gadgets and info. available.
Oh, and this line from the FA is priceless: "...is connected to the internet through advanced computer technologies such as computer clusters and GRID network." Don't get too technical on me...
Does anybody have a mirror?
It keeps changing on me!
Atmospheric noise
Lava lamps
Radioactive decay
Entropy
Call me paranoid, but I think I'd rather use a local pseudo random number generator than an external true random generator. My security concerns associated with using a local pseudo random number generator are outweighed by my privacy concerns of contacting a third party every time I want to establish a SSH connection or use my credit card online.
Great for research though, of course.
I've been waiting on this for a long time.
--- JurassicPizza
Look at the signup page. You not only need to prove that you are a human but also that you have elementary knowledge of calculus.
Save the bandwidth. Don't use sigs!
These things usually use a natural noise source as a random seed and then mathematically normalize them to produce random numbers that are useful. It's pretty standard stuff.
"Prefiero morir de pie que vivir siempre arrodillado!"
I keep getting 09f911029d74e35bd84156c5635688c0...
Beware of MITM attacks!
I wonder if I use bold in my signature, people will notice my posts.
That joke would have been a lot funnier if you had cited your source (which, by the way, is required by xkcd's cc license).
- RG>
Hey pal, this isn't a pleasantforest, so don't waste my time with pleasantries!
Apathy:
Use 123, what does it matter anyway.
For scientific research, there's a very good reason to use pseudo-random numbers: reproducibility.
If you're analyzing a stochastic model, you want to be able to generate lots of runs with different random sequences and gather statistics from the ensemble. But if you see interesting behavior in a particular run and want to take a closer look, you want to be able to go back and run it again, exactly as it happened the first time. In this case, you don't want real randomness, you want pseudo-randomness with good statistical properties. I'm currently checking through my code to make sure you can do just that when using this tool.
"Random" is a word used when an event has too many unknowns to reasonably no the outcome.
To use a very simple random event: Flipping a coin.
If you know all the variables, you will know what the outcome will be.
How heavy is the coin? what side is up at the moment of the flip? whats the air density? how hard was it flipped? etc. . .
The Kruger Dunning explains most post on
http://xkcd.com/c221.html
Great, now you're going to get yourself sued by the MPAA for randomly guessing their new encryption key!
lavarand
A similar LGPL implementation: LavaRnd
You are being MICROattacked, from various angles, in a SOFT manner.
[Only God] can generate "truly" random numbers.
Oh yeah? Being omniscient, wouldn't he know what number he'd generate before he generated it? Not too random, is it?
If you need to repeat the random series, why don't you just store the numbers in a file?
Because with pseudo-random seeds, I do. I store the 1000 seeds and run it 10,000 iterations on each run. If I were to store each random number, I'd have to store 10,000,000 numbers in my file rather than 1000. I'll always store them, but the question is whether it takes 1000 records or 10,000,000. For academic purposes, the results aren't statistically different, so why store more numbers?
Learn to love Alaska
You don't save them for the same reason that you don't necessarily save the state of the simulation after every state change. You're generating a lot of these things, conceivably billions and billions or more. Even if you do save that kind of data for the short term--be it the pseudorandom sequence or state changes or both--you may want to eventually delete it while still having the option of re-creating it at a later date. With a reproducible pseudo-random generator, you can do this by saving only the parameter settings and the seed value.
I just checksum my Windows registry.
It's random enough for my purposes.
http://web.archive.org/web/20011027002011/http://d ilbert.com/comics/dilbert/archive/images/dilbert20 01182781025.gif
[Tour of Accounting]
Accounting Troll: "Over here we have our random number generator"
Number Generator Troll: "Nine Nine Nine Nine Nine Nine"
Dilbert: "Are you sure that's random?"
Accounting Troll: "That's the problem with randomness: you can never be sure"
Not really. Ever since Einstein's theory of stimulated emission, physicists have known that some events are really and truly random, and that it is not possible, even in theory, to predict when they will occur. A good example is radioactive decay. Some atoms, at exactly the same energy levels, will decay in multiple different ways, at a time that can not be predicted.
Quantum scattering (i.e., what happens when really really small particles bounce off of each other) is also not a deterministic process: Identical particles, in exactly the same energy state, will often scatter differently. Looking at particles coming out of a Bose-Einstein condensate is a good example: they are all in exactly the ground state, but come out in random ways.
Yeah, it's weird, and Einstein never got used to it, but his big paradox about it, the EPR paradox, has been shown to physically come down against determinism and hidden variables. So, like it or not, randomness is an inherent part of nature. Deal with it.
----- Why sig when you can sign? PGP key id 7675D05E
As Kittel and Kroemer put it, "The probability of Hamlet is therefore zero in any operational sense of an event...", and the statement that the monkeys must eventually succeed "gives a misleading conclusion about very, very large numbers."
So in other words there really is *no* hope that web 2.0 will actually produce anything truly outstanding?
In the free world the media isn't government run; the government is media run.
My number's bigger than your number...
No sig today...
So, what is so hard about generating "truly random" numbers in a computer? I would think that a fairly simple grabbing of arbitrary chunks of memory and using the bytes as floats would produce something that's truly random. So long as your selection method for the bytes was sufficiently arbitrary, there should be no opportunity for pattern development. What am I missing?
The Spoon
Updated 6/28/2011
The intuitive disconnect here is that humans have a well-documented inability to understand or true randomness. By true randomness, I mean in a mathematical sense: uniformly distributed values over some range, with each value independent from the next (uncorrelated). Just try it: whatever you come up with, whether it be some algorithm, hardware, whatever — it will probably fail all the statistical tests for true randomness. And if it can't pass those tests, then it will be useless for most of the applications that one needs random things for... simulations, encryption, authentication, etc. And mind you, even if one can pass all the existing tests with some method, it doesn't mean that someone will invent a new statistical test tomorrow that the method will fail spectacularly. Like many apparently simple problems, random number generation is surprisingly deep and very, very difficult to do right.
[B.v.L]
> What am I missing?
I'll start to list some of the things you're assuming, that are wrong:
(1) The probability distribution of the individual bytes of the memory of your computer is uniform;
(2) The individual bytes of the memory of your computer are independently distributed (i.e., there is no correlation whatsoever between the value of a byte and the values of the surrounding bytes);
(3) If you could invent a "sufficiently arbitrary" method for selecting the bytes, it wouldn't already pay to use the address itself as the random number;
(4) If you construct a float from independently uniformly distributed bytes, the distribution of the value of the float would be the distribution you are interested in (see IEEE_754);
On the other hand, you are correct that this announcement is mostly a big yawn because:
(1) High quality truly random numbers have been available via the Internet a long time (just not from direct quantum sources)
(2) Hardware for generating large quantities of truly random numbers from various noise sources has been available for a long time
(3) Hardware for generating large quantities of truly random numbers directly from quantum sources is now available (Quantis)
(4) A DIY solution for (2) was posted on Slashdot recently
No, that's not what I'm saying. I'm saying that experiments have been set up that have put particles into identical states, and then watched them evolve differently in time.
When I say identical states, I mean identical states. In quantum mechanics, there is an underlying symmetry for exchange of particles. (Integer spin has no change in the wave function at all on particle exchange ==> Bose-Einstein statistics. Half-Integer spin has an inversion in the wave function on particle exchange ==> Fermi-Dirac statistics.) This is not marginal stuff -- it explains why lasers work (integer spin) and atoms have electrons with different energy levels (half integer spin). If we take an atom like Helium (even spin), and do experiments with it, we get the results from Bose-Einstein statistics, which means that the particles must be identical, since otherwise, we couldn't see those results.
I think you are underestimating physics. In physics, we don't talk about something if we can't define what it means. The gold standard in physics is a tested prediction. Almost as good is the testable prediction. The minimum level for something to be called physics (rather than theology) is the thought experiment, which might someday be turned into a real experimental prediction. An affirmative statement in physics requires experimental proof, and disagreeing with an affirmative statement takes at least a prediction why. Gut feelings don't count, and I've got a feeling (this is slashdot, not physics, so I can say that) that a gut feeling is all you have.
----- Why sig when you can sign? PGP key id 7675D05E
These are all pure mathematical algorithms. Nowhere in any of these is there any sort of pre-generated random lookup tables. (Unless you count the S-boxes used in some block ciphers with Fortuna.) Pre-generated "random" lookup tables only hide poor randomness in the generation process and don't actually improve the situation cryptographically at all; I suspect that for most other applications there would be problems as well. If your generated numbers don't cover the entire domain space uniformly, then they still won't no matter how many lookup tables you use to transform them.
According to the article, people are sitting around rolling dice to generate random number sequences. Really? REALLY?!? Who wrote this article?
[BvL]
So long as your selection method for the bytes was sufficiently arbitrary
As long as it was fairly random, one might say...
You see the problem?
Get your own free personal location tracker
Extreme Programming - Redundant Array of Inexpensive Developers
:(){
Geeks rejoice - now you can prove that was a natural 20!
Because then your own psychology comes into play.
If you ask people to pick a number between 1 and 10, the vast majority of them won't pick 1 or 10. People just don't like the edges. I think that they avoid 5, too, because it's right smack in the middle. For a number between 1 and 10 to be random, most people subconsciously want to make it not stand out and will pick something like 3, 6, or 8, thus not making it even random enough for gaming.
Also, in the same vein of not standing out, if you ask people to pick multiple numbers between 1 and 10, most won't allow there to be any patterns in them in the attempt to make them more random, thus actually making them less random. For example, if you ask people to pick five numbers, most won't pick something like 4, 4, 4, 4, and 4, even though it's a legitimate combination that's just as likely as something like 7, 3, 1, 1, 9.
Another example. When I was in high school, I used to play $5 in the lottery once a week, figuring that it sure would be convenient if I never had to bother going to college and get a job and so on. I usually just selected the quick pick and let the machine pick my numbers. Once, though, I manually picked 1, 2, 3, 4, 5, and 6 for the first ticket, 7, 8, 9, 10, 11, 12 for the second, and so on. My dad basically said, "You're an idiot. Those numbers will never come up, and you just wasted five dollars!" He never quite got it that, aside from the lottery being a colossal waste of money to begin with, it didn't matter what numbers I picked; any given set was just as likely to come up as any other given set. Not having six consecutive numbers is merely imposing human psychology on the random numbers, which could have very well been consecutive numbers.
If I'm not mistaken, several years ago, someone proved that the digits of pi are random. That if you expand it out to a bazillion decimal places, you'll eventually run across patterns like 0123456789 and such. As humans, with brains that are designed to seek out patterns, it strikes us as interesting, perhaps even as some sort of sign that the numbers aren't random. Nothing is further from the truth, though; the lack of such patterns would be a sure sign that the numbers aren't random.
Fourmilab's been doing this for years with HotBits. I remember writing an atomic-powered band name generator that used it.
Because I got this:
"Quantum Random Bit Generator Service: Sign up failed
Congratulations! You have successfully registered for QRBG Service.
Now, you can log in and check your quota and usage statistics, or just start using the Service."
I guess I have to stuff a cat in a box to see if my account actually works now.