Archive for open source

run my code [guest post]

Posted in Statistics with tags , , , , on July 18, 2012 by xi'an

(This guest post has been written by Nicolas Chopin.)

I have been contacted by Christophe Pérignon, a prof. of Finance at HEC and co-founder of, a very interesting initiative that deserves to be publicised widely. Essentially, it’s arxiv for scientific code. You can create a “companion web-site” for each of your projects, post your code (with links to the corresponding paper), and let users run your code in the “cloud”, with their own data. All of this through a simple web-page interface.

I’ve not tried it yet, and I still wonder if this does not sound too good to be true; for instance, I wonder what happens if too many people post computer-intensive programs that take hours to complete. But Christophe tells me there is some badass hardware behind the project (a big server from CNRS); they are also backed by prestigious institutions (Columbia, NSF, CNRS, etc.).

But certainly the idea is excellent, and looks like the next step in reproducible research. (One of the co-founders is Victoria Stodden, an Assistant prof in stat at Columbia, and a well-known advocate of reproducible research and open research.) One could also use it to illustrate an idea in a conference, or during a course.

The project was started by people in Economics and Business. There are still some reference to this on the web site, and indirectly through the list of currently implemented languages (Matlab, R, and … Rats!). But Christophe tells me that they want to reach further. They have already projects in image analysis for instance. They are apparently open to other computer languages (e.g. Python), if there is some demand.

It is going to be really interesting to see how much this project is going to gather steam in our field and beyond. Perhaps this is the start of a new trend where we will run more and more our programs “in the cloud”, with the added benefits of openness and simplicity. We live in exciting times!

%d bloggers like this: