I had no idea what I and Heiko would get ourselves and Shogun into, when we agreed to attend the NumFOCUS Sustainability Workshop. The stated goal of the Workshop was to “guide project leads and core contributors in identifying an appropriate sustainability plan for their project as well as the initial steps required to start implementing that plan”. Sounds great. But how to have a discussion with more than 50 people (without drifting into unproductive chaos that is), and why should I spend 14 hours on a plane to do so?
Christie Koehler, Projects Director at NumFOCUS, used a method aptly named “Future Search” to structure the discussion. In principle, Future Search works like this: reflect on the past to see where we’re coming from; take record of what’s happening in the present; identify common ground, i.e. things we all want in a bright future; and finally come up with action items how to get to there. In practice, we formed small groups of people with either similar or diverse roles in their open source projects (as part of which I found myself Shogun’s “community manager”). For each part of the framework — past, present, future and action plans — all groups discussed or worked on short tasks before presenting to everyone else. The small group discussions with other open source enthusiasts were exciting and surprisingly productive; even more the many personal conversations I had on the side; they filled the abstract concept of ‘sustainability in open source’ with a meaning: how to make our projects live long and prosper.
To me, the most striking thing about open source and its people was how much we share. Above all, we share life experiences — of studying, learning coding skills, joining an open source project, progressing in our (academic) career, assuming responsibility in the project — as one participant phrased it: “We’re past and future versions of each other”.
Also, all projects — no matter how young or mature, funded or unfunded — have the same problems: it’s hard to recruit long-term developers, no-one has enough time (or money!), the website is underwhelming (sounds familiar?), the project relies on a small number of crucial people and so on. The latter is the infamous “bus factor” — how many contributors can get hit by a bus, before project dies. Shogun’s bus number might be just about a dire 1.5.
However, the projects don’t just share problems, they also share enthusiasm, a welcoming culture, idealism, the love for the coding, snobism towards other programming languages (Shogun team: “Untyped languages suck!”), an appreciation of awkward jokes (tasters? “SunPy — AstroPy with more Photons”; “Have you considered re-naming your ‘Showroom’ ‘Shoroom’?”), and above all a passion for open source.
Together, we identified trends and external forces that currently influence all open source communities, such as: the rise of big data and cloud computing, increasing professionalisation of open source projects, demand for non-code contributions, increasing personal conflicts between job and open source project as people progress in their career, etc etc etc. We then imagined a bright future for our open source scientific computing communities and used this as a basis for short- and long-term action plans: stay tuned for the NumFOCUS report (I was too wrapped up in discussions to take notes. Sorry.)
In short, I learned a ton, my contributions were valued and I had fun. Importantly, the workshop led me and Heiko to brainstorm how this all relates to Shogun. We can’t wait to have these discussions with all of Shogun’s contributors over the next months and during our upcoming Hackathon.
