Sat, 27 Oct 2007
Where's that blog?
I said a while back that I would try to publish more regularly, and not wait until every article was perfect. But I don't want to publish the unfinished articles yet. So I thought instead I'd publish a short summary of what I've been thinking about lately.
I hope to get at least one or two of these done by the end of the month.
Simplified PokerI recently played a computer poker game that uses a 24-card deck, with only the nine through ace of each suit. This changes the game drastically. For example, a flush is less likely than a four of a kind. (The game uses the standard hand rankings anyway.) It is very easy to compute optimal strategies for this game, because there are so few possible hands (42,504) that you can brute-force all the calculations with a computer.
This got me thinking again of something I started writing up last year and never finished: The game of "Simplified Poker", which was an attempt to do for Poker what the λ-calculus does for computation: the simplest possible model that nevertheless captures all the essential features of the original. Simplified Poker is played with an infinite deck in which half the cards are kings and half are jacks. Each hand contains only two cards. Nevertheless, bluffing is still possible.
The Annoying Boxes PuzzleThis is a logic puzzle in which you deduce which box contains the treasure, but with a twist. I thought it up many years ago, and then in the course of trying to write up an explanation about five years ago, I consulted Raymond Smullyan's book What is the Name of This Book? in order to get a citation to prove a certain fact about the form that such puzzles usually take. In doing so, I discovered that Smullyan actually presented the annoying boxes puzzle (in slightly different form) in that book!
It's primarily waiting for me to take a photograph to accompany the puzzle.
Undefined behaviorI have a pretty interesting article on the concept of "undefined behavior", which is a big deal in the C world, but which means something rather different, and is much less important, in Perl.
[ Addendum 20071029: This is ready now. ]
TootleMy daughter Katara has become interested in the book Tootle, by Gertrude Crampton, which is the third-best-selling hardback children's book of all time. A few years back I wrote some brief literary criticism of Tootle, which I included when I wrote the Wikipedia article about the book. This criticism was quite rightly deleted later on, as uncited original research. It needs a new home, and that home is obviously here.
Periodicity without Fourier SeriesSuppose I have tabulated the number of blog posts I made every day for two years. I want to find if there is any discernible periodicity to this data. Do I tend to post in 26-day cycles, for example?
One way to do this is to take the Fourier transform of the data. For various reasons, I don't like this technique, and I'm trying to invent something new. I think I have what I want, although it took several tries to find it. Unfortunately, the blog posting data shows no periodicity whatsoever.
Emacs and auto-mode-alistThe elisp code I've been using for the past fifteen years to set the default mode for Perl editing in Emacs broke last week. My search for a replacement turned up some very bizarre advice on IRC.
Van der Waerden's problemAlso still pending is the rest of my van der Waerden problem series. I have written about four programs so far, and I have two to go.