Wednesday, September 07, 2011

Somebody's Doing Virtual Shakespeare Monkeys (Again)

I almost missed it, but looks like the idea of writing computer code to simulate the "infinite monkeys typing Shakespeare" thing has come around again.

As a programmer, I know what it's like to have free time and to do stuff like this (well, I used to :)).  So we won't debate the uselessness, or the apparent misunderstanding between "infinite" and "millions and millions."  You can have all the millions you want, it's still not infinite.

Instead I'll just point out that it's been done before. The source link is long dead, but my blog post back in 2005 should serve as proof that the idea is far from a new one.

When is somebody going to try the Schroedinger's Cat experiment, and give the monkeys a rest?


kj said...

So there's a brand-new Shakespeare play—only it's in a box. And you don't know whether it's a comedy, history, tragedy, or romance until you open the box.

And if the cat is dead, it's a tragedy. If it's married, it's a comedy.

And if it doesn't fit in any other category, it's in some other category.


Jesse Anderson said...

I understand the less than infinite nature of the project. If you infinite monkeys could you send half my way?

BTW I checked your post on the other project that did this. Getting 26 character groups from data is very lucky. This would be go get some lottery tickets lucky. The odds of getting a random hit on Shakespeare is ~3 million in 26^26 or 6 followed by 36 zeros. I am guessing there were some bug or a misunderstanding in that program.

Duane said...

Hi Jesse! Always fun when the author of the source material stops by :). I'm curious whether you spotted my link in your referrer logs so quickly, or if you were actually following us already?

On a different note, don't you think that your denial of the other project getting 26 letters is detrimental to your project as well? You're basically assuming that you'll never get a meaningful result, because if you do, then some other programmer will come along and say "Yeah, that's unlikely, that's probably a bug."


My point about the infinite thing was always that, by definition, if you had an infinite number of random character generators (let's call them what they are), then all possible combinations of random characters of any given length would be represented. Therefore, one of them would contain the complete works.

Jesse Anderson said...

Yes, I came through the referrer in Google Analytics.

I don't think the difference is detrimental to my experiment. I actually ran back of envelope and empirical checks to make sure this my 9 character sets were doable in a non-infinite amount of time. The exponential growth of the problem led me to take some liberties with how the problem is run through. Getting several 26 character hits is pretty unlikely (but possible) because I actually did the one too and didn't get a single hit.