- 09:30–11:00: lecture
- 11:00–11:30: coffee break
- 11:30–13:00: lecture
- 13:00–14:00: lunch
- 14:00–15:30: lecture
- 15:30–16:00: coffee break
- 16:00–17:30: lecture
The hype about reproducible science seems to imply that putting on GitHub the scripts that re-create the plots of your paper is all that is needed. Have you ever tried to use code published by other scientists to do anything else than reproduce their plots? Have you been forced to learn how to manage docker containers on the institute's cluster because the scientific software you wanted to use could not be installed in any more useful way? Did you have to spend precious months of your PhD by trying to figure out what the code of that student who left the lab was doing? Did you have to re-write some of your code from scratch because you could not understand it anymore?
The goal of this lecture is to give you the minimal infrastructure needed to
make your code re-usable by others (and by your future self). You'll learn how
to organize your code following well established standards, how to publish your
code so that it can be installed by a simple pip install, how to make it
robust by linking testing with continuous integration, and, if there is time,
how to document it properly.