Bitcoin Forum

Bitcoin => Bitcoin Discussion => Topic started by: alan2here on August 13, 2013, 04:20:29 PM



Title: Random Value Pedantry
Post by: alan2here on August 13, 2013, 04:20:29 PM
I've seen some talk about this recently, I thought I better check my assertions.

Also, this is not about the recent Android wallet issues, which was about random numbers, but was a bug.

Take 3 available functions, they all on request within 1/100th of a second generate 64 random bits.

A: Uses a standard common generator.
B: Uses one of the best open source cryptographically oriented generators instructed to take all allocated time to generate the value.
C: Uses a web service that receives values from decay of an alpha source.

Can I tell to a statistically significant degree, which generated a given group of bits?

Manually? No
Using any available software? I don't know.
Using any available software between B and C? I imagine not.

So, even if definitive proof exists/arrises that C uses a fundamentally random process, is it ever relevant that B is only 'sudo' random?


Title: Re: Random Value Pedantry
Post by: EmperorBob on August 13, 2013, 09:05:49 PM
http://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator (http://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator) is what you're looking for.

By definition, a good pseudorandom number generator is indifferentiable from a true random source.