Talk:Random number generator

Paging a real mathematician...paging a real mathematician...

I don't have ready access to the SIAM Journal for Blum-Blum-Shub, and it's been a while since I've even formatted the equations. Collaboration here would be very very welcome, but what we need is an article that has useful content, not just links and editorializing.

Howard C. Berkowitz 15:26, 3 August 2008 (CDT)

Assuming you are talking about "A Simple Unpredictable Pseudo-Random Number Generator", I have access to the article. What do you need to know? I know very little probability theory and statistics though, so I'm not sure how much help I can be. -- Jitse Niesen 04:27, 4 August 2008 (CDT)

That's it, I believe. They aren't my strongest mathematical areas, but I wanted to see a little more of what the paper actually said, and get an idea how it would be implemented, than the summaries/reviews. I didn't follow one proof appearing elsewhere, and hoped the original might be better written. Howard C. Berkowitz 11:03, 4 August 2008 (CDT)

Structure — how many articles

Wikipedia has [1] as a disambiguation page with links to several other things. Do we need that structure or something similar here? Certainly pseudo-random generators and true RNGs are quite different. How many articles do we need and with what structure? Sandy Harris 16:15, 3 August 2008 (CDT)

It's not clear to me that Wikipedia is setting tbe best example. In this particular case, the article is titled "Random number". Physical-measurement-based and pseudorandom generators are subtopics. The techniques for generating random sequences that follow a probability density function are common to physical-random and pseudo-random generators. The techniques of testing also apply to both, as it has been shown that apparent physical sources of randomness may show self-similarity. When the self-similarity is demonstrated, post-processing can break it and give a non-self-similar output.

This structure, I believe, works without breaking up the core concepts. It would be perfectly reasonable to have articles on specific techniques, such as BBS or radioactive counting.

In other words, what is the problem that needs to be solved? Howard C. Berkowitz 19:39, 3 August 2008 (CDT)

I'm not sure there is a problem. However, in case there might be one, it seems worth discussing early on to ensure we get the structure we need right before doing a lot of writing.

For example, I'd say pseudo-random generators and real random number generators need to be separate articles (of course linking to each other). WP has articles for both, and another separate one on applications of randomness. I'm not sure of the right structure, but I think some discussion is called for. Sandy Harris 20:01, 3 August 2008 (CDT)

Ummm...might I suggest that citing Wikipedia as the reason for doing something might not be the best argument on Citizendium, which was established as an alternative to Wikipedia, and many authors and editors came to CZ to get away from Wikipedia?

RFC 4098 discusses physical and pseudorandom generators in an integrated way. Knuth starts out discussing PRNGs, but the rest of the chapter (about half the book) is about testing for randomness, regardless of the source. Schneier, in section 17.13, discusses PRNGs and physical generators together. Goldwasser and Bellare discuss, in the same chapter, PRNGs, removing self-similarities from physical sources, and proofs that some PRNGs may be theoretically, but not computationally, breakable.

So, no, I don't see any good reason to separate the two areas. In fact, I think it's highly advisable to discuss the general approaches in a compare-and-contrast way, although it is very reasonable to have separate articles on detailed individual methods. Howard C. Berkowitz 20:15, 3 August 2008 (CDT)