Comments on: Quasi-random sequences in art and integration

By: Rick Wicklin

Rick Wicklin — Thu, 27 Oct 2011 19:31:31 +0000

I reached similar conclusions, not in art or integration, but on the theatrical stage. Turns out that when directors tell cast members to “go to a random position” on stage, they really mean “go to a quasi-random position.” For details, see
http://blogs.sas.com/content/iml/2011/01/28/random-uniform-versus-uniformly-spaced-applying-statistics-to-show-choir/

By: John

John — Thu, 15 Apr 2010 22:00:12 +0000

Quasi-random sequences have aesthetic applications, but they can be objectively defined by their mathematical properties. For this reason they are sometimes called “low discrepancy sequences” where “discrepancy” has a technical definition.

By: CogitoErgoCogitoSum

CogitoErgoCogitoSum — Thu, 15 Apr 2010 21:00:46 +0000

Another realization I had. How do we know that quasi-random is merely appealing to human senses. How do we know that it isnt more true to randomness than what you deem “random”?

These pictures with the dots… they were generated on a computer, right? The regular “random” one wasnt truly random to begin with… it was pseudo-random, generated by a computer.

My observation is just that, which we deem random and which quasi-random is arbitrary, isnt it?

Could it be that the inherent “clumpiness” of “random” events is in fact the inherent pattern to be found in all pseudo-random generating algorithms? And that the unappealing nature of pseudo-random “randomness” is a legitimate concern?

Of course Im not suggesting that quasi-random is any less pseudo-random on a computer, but perhaps the algorithm has been adjusted, the outcomes modified, to be truer to actual random than you would like to believe. You presume that quasi-random only “looks” more random, and that clumpiness is actual random. But Im wondering if you have it right, if you just dont have too much faith in the computer.

By: John

John — Wed, 14 Apr 2010 01:09:49 +0000

There are goodness of fit tests that will distinguish quasi-random sequences from random sequences.

By: CogitoErgoCogitoSum

CogitoErgoCogitoSum — Wed, 14 Apr 2010 00:56:05 +0000

Here is another question for you… a good one…

You say that there is a distinct difference between “true” randomness and what humans “feel” to be random. The latter being described as “quasi-random”.

Is it possible, then, to analyse a seemingly random set of data… and discern how random it is. To assess, based on “clumpage” and lack of “clumpage”, and decide whether or not the event truly was random… or if a human being tried to make it look random.

By: CogitoErgoCogitoSum

CogitoErgoCogitoSum — Tue, 13 Apr 2010 18:52:57 +0000

Can we assess the quality / validity of the conclusions drawn from all laboratory experiments performed to date, which relied on a version of randomness, if we decide now that the randomness they used wasnt good enough? Exactly how does this new realization of quasi-randomness and the long established use of pseudo-randomness reflect / affect the quality of experimentation and the validity of its conclusions? Is it fair to look back on the history of science with doubtful contemplation?

By: Mark Dewing

Mark Dewing — Wed, 18 Mar 2009 04:26:10 +0000

Here's a presentation from Ken Hanson describing other methods for generating sets of points. It starts off with digital halftoning algorithms and moves on from there.

By: Jason Dyer

Jason Dyer — Tue, 17 Mar 2009 23:25:39 +0000

Here’s a fun post from a different blog (see the second footnote) which mentions trying to fill a screen with a flawed random number generator on the IBM PCJr:

http://www.wurb.com/stack/archives/530

By: Thomas Guest

Thomas Guest — Tue, 17 Mar 2009 08:30:40 +0000

Thanks for this fascinating summary, John. Without really knowing it, I've been aiming for some quasi-random aesthetic effects. See the graphics here, for example, which I created by programatically nudging coordinates away from grid squares (using a random number generator). Maybe I should use a proper quasi-random sequence?