SoylentNews

"exec" writes:

Arthur T Knackerbracket has found the following story:

"As an instrument for selecting at random, I have found nothing superior to dice," wrote statistician Francis Galton in an 1890 issue of Nature. "When they are shaken and tossed in a basket, they hurtle so variously against one another and against the ribs of the basket-work that they tumble wildly about, and their positions at the outset afford no perceptible clue to what they will be even after a single good shake and toss."

How can we generate a uniform sequence of random numbers? The randomness so beautifully and abundantly generated by nature has not always been easy to extract and quantify. The oldest known dice (4-sided) were discovered in a 24th century B.C. tomb in the Middle East. More recently, around 1100 B.C. in China, turtle shells were heated with a poker until they cracked at random, and a fortune teller would interpret the cracks. Centuries after that, I Ching hexagrams for fortunetelling were generated with 49 yarrow stalks laid out on a table and divided several times, with results similar to performing coin tosses.

But by the mid-1940s, the modern world demanded a lot more random numbers than dice or yarrow stalks could offer. RAND Corporation created a machine that would generate numbers using a random pulse generator. They ran it for a while and gathered the results into a book titled A Million Random Digits with 100,000 Normal Deviates. What now might seem like an absurd art project was, back then, a breakthrough. For the first time, a nice long sequence of high-quality random numbers was made available to the public. The book was reprinted by RAND in 2001 and is available on Amazon.

A similar machine, ERNIE, built in Bletchley Park in the 1940s for WWII, was reused after the war to generate random numbers for the UK Premium Bond lottery. To quell fears about the fairness and accuracy of ERNIE, the Post Office made a great documentary called The Importance of Being E.R.N.I.E. It's worth a look:

In 1953, randomness was finally formalized into a real computer, the Ferranti Mark 1, which shipped with a built-in random number instruction that could generate 20 random bits at a time using electrical noise. The feature was designed by Alan Turing. Christopher Strachey put it to good use by coding up a randomized love note generator. Here's an sample love note, from David Link's 2009 resurrection of the program:

But Turing's random number instruction was maddening for programmers at the time because it created too much uncertainty in an environment that was already so unpredictable. We expect consistency from our software, but programs that used the instruction could never be run in any consistently repeatable way, which made them nearly impossible to test.

What if a random number generator could be expressed as a deterministic function? What if it could be called repeatedly to deliver a sequence of random numbers, but under the same initial conditions it would always produce the same sequence? Enter the pseudorandom number generator (PRNG).

-- submitted from IRC

Original Submission