Close Cousins

Sun, 30 Oct 2016 00:00:00 +0000

This post originally appeared on the Software Carpentry website.

Our process for developing and maintaining lessons has grown and changed over time. Simultaneously but separately, an organization called the Programming Historian has crafted a diverse set of open, reusable lessons on computing skills for people working in the digital humanities (DH), and their process is different from ours in some interesting ways.

The main elements of our approach are:

A first version is created by:
- someone writing something on their own (or translating something they’ve written before),
- a group of people getting together at a hackathon to create a roadmap, or
- someone driving an open design process of the kind used for our new Python lesson or Zack Brym’s new visualization lesson.
The lesson is put in a GitHub repository. Everyone is invited to submit enhancements and changes by filing issues and/or submitting pull requests, and to comment on other people’s submissions. This is a doubly-open process: both the submissions and the reviews are tied to the GitHub usernames of their creators, which in turn are usually tied to their actual identities.
One or two people act as the lesson’s maintainers (a term we borrowed from open source software projects). Their role is primarily editorial: they either review PRs themselves or make sure that other people review them, and have final say over whether changes are merged or not.
We publish our lessons twice a year by tidying them up and then archiving them at Zenodo, which gives each of them a DOI. Everyone whose work has been merged into the lesson is listed as a contributor, and the maintainers are listed as editors (because that’s a role everyone in academia understands).

The strengths of this approach are that the community maintains the lessons (we’ve had about 400 distinct contributors in the past three years), while the editor-vs-contributor distinction allows us to recognize people who are doing extra work. Its weaknesses are that big changes are more difficult to make than they would be if there was a single author, and there’s no incentive for people to do reviews: someone’s name doesn’t show up in the bibliographic record for a lesson if “all” they did was craft hundreds of lines of thoughtful feedback.

Programming Historian on The Carpentries

Close Cousins