Comments on: Garfield + Markov Chain = Garkov

By: Homer J. Simpson

Homer J. Simpson — Mon, 09 Jun 2008 19:26:37 +0000

I wonder what it is about Garfield that has inspired people to do so many weird things with Garfield strips. In any case, I find this phenomenon pretty interesting. It’s probably been many years since I paid much attention to Garfield, but now I think I’m a fan again.

By: twodollars

twodollars — Mon, 09 Jun 2008 03:38:10 +0000

very awesome. In my 10 minutes of messing around, I came across this

http://joshmillard.com/garkov/showstrip.cgi?strip=32730

By: InSpain

InSpain — Mon, 09 Jun 2008 03:11:50 +0000

In Spain we hate that stupid orange cat.

By: jmp478

jmp478 — Sun, 08 Jun 2008 22:23:46 +0000

Wow, computers suck at jokes.
Better leave it to Jim Davis

By: ted

ted — Sun, 08 Jun 2008 21:53:45 +0000

I guess it goes to show how useful math really is.

Don’t get it. Don’t care to get it.

By: Chaz

Chaz — Sun, 08 Jun 2008 20:59:13 +0000

I like http://joshmillard.com/garkov/showstrip.cgi?strip=15733

By: v.dog

v.dog — Sun, 08 Jun 2008 20:17:33 +0000

Yeah, most of them make no sense at at all, but occasionally you find a gem like this one:
http://joshmillard.com/garkov/savestrip.cgi?strip=14226

By: Josh Millard

Josh Millard — Sun, 08 Jun 2008 19:39:28 +0000

Hi! Thanks for the link! I’m glad you’re liking it.

And agreed, CheeseDuck: most of the generated strips are just plain nonsensical. Pure Markov models are really naive things — that they can generate readable text at all says more about the structural redundancy of natural language than anything. But it’s that weird, fractured insensisibility that I find so charming about the output.

By: CheeseDuck

CheeseDuck — Sun, 08 Jun 2008 19:14:57 +0000

They don’t make any sense