Writing nondeterministic software also requires a safe and demonstrable source of randomness. But building a truly random number generator is more difficult than you might think.
Pseudorandom generators aren't perfect, they contain repetitions and regularities. And even generators based on the observation of radioactive decays or quantum events can easily suffer from "biasing" caused by the inaccuracies and "memory" of the macroscopic components that are used to observe them..
Considerations on Random Number Generators
– TrueRandomNumberGenerators_ITA.pdf
– TrueRandomNumberGenerators_ENG.pdf
Generation of random numbers with quantum techniques and QSharp applications
– www.unipd.it/news/progettoquasar
– docs.microsoft.com quantum-random-number-generator with qsharp
– docs.microsoft.com develop-quantum-applications-locally
– www.fourmilab.ch/hotbits
– www.fourmilab.ch/hotbits/hardware3.html
Check that the data is really random
– www.fourmilab.ch/random
Sites that provide quantum random numbers
– www.randomnumbers.info
– www.fourmilab.ch/hotbits/secure_generate.html
– qrng.anu.edu.au
– random.org (based on atmospheric noise)