29 Aug 2011

 I just spent a weekend in Paris, to attend the fourth European Scientific Python conference.

  • A very nice and talented crowd of roughly 160 people. It was very well-organised, many thanks to the team who made it happen.
  • Python has come a long way. I learned that Python detected the latest, scientifically sensational, Supernova explosion and will command an on-board camera in a robot on Mars in 2016.
  • "Mr. ipython", Fernando Perez, gave an insightful keynote talk. He showcased a new kernel-clients model in the upcoming ipython 0.1.1. Not only can several clients work on the same ipython process, a client can now be anything ipython users ever dreamed of (or didn't even know they should be).  First, Fernando showed a console emulation window written in QT. It had nice syntax highlighting, tex support, actually working multi-line editing and inline plots. Pretty neat. Then he showed a broser implementation, merged last tuesday, which they call the ipython notebook. Basically, the ipython session becomes a document, which is editable inline and at all times and can contain many content types (he easily inserted images, videos, math and javascript). The whole book can be saved as json. Fernando closes with the sad fact that less than 30 people do 80% of the work in many python scientific computing libraries (scipy, matplotlib, numpy, etc).
  • I presented my simulation framework Nicessa on a poster and decided to give a short lightning talk about this vision of independent layers which seem to evolve in this community (Nicessa can be the middle layer):
  • On the train ride back, I happened to sit next to Ralf Gommers, one of the few major SciPy contributors, by coincidence. Pretty interesting, not only because I actually started using SciPy in a research project very recently.
# lastedited 29 Aug 2011
You are seeing a selection of all entries on this page. See all there are.