Random Reading

Perhaps I was inspired by the idea of random in our recent discussion (okay, monologue) about random noise, but I thought of one more little bit of random noise to finish it off.

Namely, I added a link, which should appear in the top of the right-side column, that, if you click on it, will produce a randomly* selected posting from this blog. With (just now) over 1,175 to select from you can waste quite a bit of time without ever having to make any volitional choices. Amazing!
———-
*Well, pseudo-random to be precise. What's the difference? Random-number generators on computers are not actually random but they do produce a sequence of numbers that appear very nearly random. You know they're not random because if you start one over in exactly the same way as before, it will produce exactly the same sequence of numbers as before; such sequences also repeat themselves sooner or later, although the better designed ones can produce sequences of 10,000 numbers or more before they repeat.

Computers running programs, which are deterministic machines (at least these days, even if some Microsoft products don't seem all that predictable), cannot produce sequences that are actually random. However, there are simple pieces of hardware that can be added to a computer system in order to generate sequences of real, actual random numbers.

In fact, I just learned from Wikipedia (Random Number Generators) that 8-bit Atari computers actually contained an electronic circuit that sampled random electronic noise to generate actual random sequences. This is discussed in an article ("RANDOM ATARI: Enhancing the number generator", by David McIntosh, dating from March 1989) that talks about how to generate a pseudo-random sequence for testing programs, when one needs reproducibility. Some people are never satisfied!

Posted on October 16, 2008 at 23.32 by jns · Permalink
In: All, Briefly Noted

2 Responses

Subscribe to comments via RSS

  1. Written by Tim Wilson
    on Saturday, 18 October 2008 at 08.41
    Permalink

    This is an issue where Steven Wolfram has some interesting ideas. While deterministic algorithms can produce quasi-random outcomes (the pseudo-random noise you just discussed), he argues, if I remember correctly, that the same deterministic cellular automata algorithm can produce results that eventually are constant, steady-state but changing (e.g., ramp or periodic), or truly random (never repeating) depending on initial conditions.

    I'm too lazy to whip out my copy of A New Kind of Science to come up with the hard references, but what he is getting at is the idea that our beloved "real" randomness — e.g., electronic thermal noise — might have underneath it a deterministic algorithm/process, just one occurring at scales we currently don't resolve/observe.

  2. Written by Tim Wilson
    on Monday, 17 November 2008 at 06.43
    Permalink

    This xkcd illustrates Wolfram's idea. Well, in a sense.

Subscribe to comments via RSS

Leave a Reply

To thwart spam, comments by new people are held for moderation; give me a bit of time and your comment will show up.

I welcome comments -- even dissent -- but I will delete without notice irrelevant, rude, psychotic, or incomprehensible comments, particularly those that I deem homophobic, unless they are amusing. The same goes for commercial comments and trackbacks. Sorry, but it's my blog and my decisions are final.