On August 22, Marty McGuire (@schmarty) and I (@joshua_r_smith) taught a for-profit workshop based on Software Carpentry materials, particularly the novice Python lesson. This post is a debrief of that workshop.
We forked the workshop-template and the python-novice-inflammation repos to build a workshop website and notes for the Python lesson. We decided to only do a one-day class because we did no marketing in the sense that we did not target any specific group for this workshop. For typical Software Carpentry workshops, part of the marketing is already done: SWC generally does workshops for some subset of science people.
We also charged $20 US per person for the workshop; this fee was to mitigate no-shows. (The fee was collected via Eventbrite.) The goal of this workshop wasn't to make money, it was to go through the process once and figure out where the bugs are. We coordinated over the phone and via the issue tracker on a private GitHub repo we set up.
All of the attendees were highly impressed by the quality of the Software Carpentry material. Many of them found it via the attribution link we put on our site according to SWC's attribution rules.
We were very grateful for the level of detail with regards to set up instructions, e.g. the operations guide. and checklists. It is clear that these documents were created at the cost of many mistakes made by prior instructors.
Finally, the liberal Creative Commons license of the SWC materials was really the enabling feature for this workshop. We are very grateful for all of the hard work SWC volunteers have donated to making this material excellent.
The "Could Be Better"
Two major things stand out, which are related.
It was more difficult than it should have been to re-brand the lesson materials to comply with SWC's copyright disclaimer:
All of our lessons are freely available under the Creative Commons - Attribution License. You may re-use and re-mix the material in any way you wish, without asking permission, provided you cite us as the original source (e.g., provide a link back to this website).
Having both the Markdown source and the HTML built from it in the same repo, in the same branch was non-optimal. This setup makes sense based on the varying levels of Git mastery of the contributors especially regarding contributors' familiarity with Git's branching and merging model. Nevertheless, it felt like it was counter to version control best practice.
These two issues could likely be mitigated by switching to a static site generator (such as Jekyll) to build the HTML pages from some Markdown source. Clearly, the Shirtable team aren't the first to have this idea. Once the transition to Jekyll has been made, the actual lesson material can live in Markdown files and whatever assets are required by the lesson (data files, images, Jupyter Notebooks, etc.). The branding could even be stored in a separate repo that's brought into the lesson repo as a submodule. I'm pretty sure that's how Pelican does it. It seems like all of these details could be handled by a build script.
Solving this issue is important, and not just to help for-profit entities like Shirtable. Anyone, even not-for-profits and educators, will need to re-brand these materials in order to comply with the rules. The Shirtable team believes the difficulty associated with rebranding is a higher barrier than necessary for educators with limited temporal resources to wrangle the source.
All in all, we accomplished the objective we set out to accomplish. The workshop instructors and participants all found the SWC Python material to be valuable, informative, and easy-to-follow.