I'd love a book on statistics for programmers written by Peter Seibel. I know there's Think Stats, and that's a pretty good book, but I'd be interested in a Lisper's take. PCL and Coders at Work are really great.
I actually have the same feeling as the previous poster. I'd buy/kickstart his stats book in a heartbeat. I don't think Lisp users would be "better" at teaching stats in some objective sense, but many of them (not all) have a certain turn of mind which looks like "clarity" through my subjective lens.
It is probably like Michael Spivak writing _Physics for Mathematicians_, because he didn't understand physics books written by physics people. PDF where he explains his troubles with elementary physics: (http://www.math.uga.edu/~shifrin/Spivak_physics.pdf)
I also loved the Berkeley book, Structure and Interpretation of Signals and Systems. Went down well with a 5 week daily diet of "Thinking in Systems: A Primer".
> what makes a lisper better at statistics teaching and exposition than a non-lisper? (say, a trained maths educator?)
Nothing really, but I think the code portions of a book like that might be interesting. Allen Downey's book has a bit too much layering of OOP for the sake of OOP for my taste. Check it out if you haven't, he builds up classes for most things where simpler data structures could be used, and ends up wrapping everything in what's effectively a custom API.
