Takeshi's Komanechi University Mathematics

I suppose some of you might already know Takeshi Kitano as a filmmaker, who shot great movies like Sonatine and Kikujiro.

Here in Japan, he is very famous, but not for his movies. More that a filmmaker, he is popular as a comedian and appears a lot on TV, know as Beat Takeshi.

But what you might not know, even living in Japan, is the Math Kitano the mathematician. Indeed, during his childhood, Takeshi's mother was very strict concerning education. He went to the renown Meiji University and usually says that he would have liked to become a mathematician if he didn't become a comedian.

This passion for mathematics gave birth in 2006 to the TV program "Takeshi's Komanechi University Mathematics" (たけしのコマネチ大学数学科).
Aired every Thursday between 25:15 and 25:45 on Fuji Television, this program aims at showing the charm of mathematics while being as well an entertainment program.
Every week, three teams are trying to solve a problem (of university entry level): Takeshi, aka "Math Kitano", two pretty student girls from Tokyo University, and a a group of four comedians.

Surprisingly, Math Kitano performs very well and sometimes outperforms Tokyo University students, especially in geometry related problems, which must be one of the reasons why the way he shots his movies is so geometric.
On the other side, the team of four comedians always use very weird but funny ways to solve problems, and it is quite enjoying to watch, even from a mathematician point of view.

After discovering this TV program, I just became addicted, and found myself solving these problems every week. I think this is one good way to keep one's mathematics abilities, and I would recommend it to any people living in Japan.

To give you a taste of it, here is a link the some problems on my Picasa galleries:

たけしのコマネチ大学数学科

And here's a link to a YouTube search results page.

I hope I will find some time solve problems in the next posts.

Think twice before naming your new programming language

Smart people invent new programming languages, with powerful features, that perform really well. What else do we need?

When you want to learn and use these new programming languages, you might want to buy books or to search for tutorials, code examples, blog articles, etc.

There comes the name: you might have plenty of good reasons for choosing a certain name for your programming language, but before deciding on it please think about people who are willing to use it.

I am one of those: recently, I am learning the Factor and J programming languages, or I should say "I am trying to learn". I suppose that you can imagine my frustration when I try to search for information using Google, or for books using Amazon.

Just think about what kind of results you will get when looking for something related to Factor on Google...

Of course you could tell me that there is the Factor documentation. But I don't want to limit myself to the official documentation: I want to read books if they exist, I want to read code snippets shared by other people, I want to read blog posts, search on GitHub (Factor is not registered yet), and the reality is that I just can't. This is so frustrating, especially for language such as J, which are even more difficult to search for, and I don't need to explain you why...

Some work have been done to make C and C++ easy to search on the internet, but nothing yet for J yet. I think the same goes for R, and if I read this wiki page, it seems that I could quote almost the whole alphabet!

So please, I beg you, the next time you create a great programming language such as Factor or J, think carefully about a name that will be easy to search for on the internet. Do not use a one-character or two-character name, only use characters from the Latin alphabet, do not use a common word, and just make it as unique as possible.

Project Euler and OCaml prime number list generator

Half a year has passed... The main reason why I didn't write anything concerning Ruby golfing is mainly because I stopped practicing quite a while ago, so I do not have any fresh tips to write about now.

Recently, I am mostly spending my time learning the programming languages Clojure and Factor, using them to solve Project Euler problems. Clojure reminds of Scheme, but I am very new to the concatenative paradigm of Factor, as I never learned Forth or any similar language.
Providing a fairly succinct syntax, they perform very well, which is the main reason why I'm now using them over Ruby.

Some weeks ago, this article in which are analyzed the results of the programming language benchmarks game made me want to use again the OCaml language, that I had previously learned and used during my studies at the university. This analysis shows it as an almost ideal programming language in terms of code size vs performance trade-off.

Several problems of the Project Euler deal with prime numbers, so I needed some code to generate them. Unfortunately, I couldn't find any simple prime generating code on the internet: it seems that finding OCaml code snippets is not an easy task...
Here is quite simple implementation of the Sieve of Eratosthenes which is quite fast (even when running only bytecode) and should be enough for most of the problems that require to use prime numbers in the int range:

Which will for example produce the following result:

# let primes = primes_upto 4000000 in
  List.nth primes ((List.length primes) - 1);;
- : int = 3999971

For problems requiring bigger primes, but not an exhaustive list of them, using algorithms such as the Miller-Rabin primality test is a better choice.