Everybody knows how important documentation is. It lets new users learn the contents of a package quickly, it lets current users quickly answer questions they have about a package, it also often gives examples of the "best" or "proper" ways to use a package.
However, as important as it is, we often don't give documentation enough attention. Docstrings are incomplete or missing argument explanations, examples don't compile or reveal outdated APIs, tutorials are often idiosynchratic (or not even there).
We think documentation is great, so we're devoting a week to improving the documentation in the open-source ecosystem. We're going to call it "the docathon".
For more information about the docathon, check out the website here
It's kind of like a hackathon, but focused on developing material and tools for documentation. This is the first of hopefully many docathons to come!
The first docathon will be tentatively held in early 2017, more details to come!
We'll have one meeting at the beginning of the week to discuss potential projects and pair interested helpers with packages / projects that need work. We'll have a physical presence at the Berkeley Institute for Data Science, but all are welcome to participate in person or remotely.
As a group, we'll spend the week working on various projects that aim to improve the documentation in our open-source ecosystem. This might be improving the docstrings of a function, or upgrading examples to using a new docs technology, or even building infrastructure and tools for improving documentation.
Check back here soon for a form that we'll use to field ideas for projects and organize people to work on them!
- matplotlib (see with NelleV)
- scikit-learn
- mne