The Software Stewardship Lab launches today, a Scottish non-profit set up to do applied research on the open source ecosystem, and I’m one of its directors. Vlad-Stefan Harbuz, who runs the Open Source Pledge, is the executive director and did the bulk of the work getting it incorporated.
I’ve been trying to do this kind of research for about a decade, starting with Libraries.io in 2015 as a side project indexing every package registry I could find, which was acquired when the hosting bills outgrew what I could cover on my own. ecosyste.ms is the second attempt at the same idea, kept running by a sequence of grants and contracts that each cover a year or two of server costs, with most of my own time on it either unpaid or subsidised by other contracting work.
The organisations set up to address open source sustainability have the same funding problem as the maintainers they are trying to help, often a worse version of it because a maintainer at least produces software that people want to use, and almost everyone I know doing this kind of research is patching together an income the same way I am. Grant programmes fund a project for a fixed term and then stop, which is also roughly how the academic contracts work for researchers studying maintainer burnout, while the tools built to measure the ecosystem need servers and bandwidth that someone has to keep paying for year after year regardless.
The funding pattern also pushes the research towards one-off snapshots rather than systems that keep collecting and reporting, so each new grant redoes much of the same data collection from scratch, and a large part of why ecosyste.ms exists is to be a continuously running layer underneath that later work can build on. I’ve watched a fair number of good efforts wind down when a grant ended or a sponsor changed priorities, and the questions they were working on are all still open.
The Lab is an attempt to give that work an institutional home with a horizon longer than a single grant cycle, structured as a research lab that contracts researchers, publishes open-access papers and open datasets, and builds tools that stay open source, on the basis that instruments for monitoring critical infrastructure are themselves critical infrastructure and shouldn’t sit behind a licence fee.
Work is also under way with solicitors to register it as a charity, which if granted would make it the first UK charity working on the sustainability of the open source ecosystem, and would mean a regulator accepting that as a charitable purpose in its own right.
The initial research themes are supply chain security, maintainer wellbeing, funding models, and governance, which overlap heavily with what I’ve been writing about here for the past year, and a chunk of that writing is already syndicated on the Lab’s site alongside work from the others.
I’ve been reading and citing the rest of the board for years, which is a large part of why I said yes: Miranda Heath, a psychologist at the University of Edinburgh, wrote the most thorough study of burnout in open source I’ve come across, and Dawn Foster brings a PhD in software metrics along with board seats at CHAOSS and OpenUK.
Daniel Roe leads Nuxt, Matias Capeletto is a core Vite developer, and the two of them are behind npmx, which I wrote about back in April, so the board includes people who maintain a sizeable share of the modern JavaScript toolchain and know first-hand what being on the receiving end of sustainability advice is like.
Mike McQuaid, who I worked with at GitHub and now maintain Homebrew alongside, is advising, and I’m lucky to be working with that group on this.
The stated long-term goal in the pitch document is to raise enough recurring funding to offer open-ended stipends to people doing this research, so the work can outlast whichever grant happens to be paying for it in a given year, and after a decade without anything like that available it’s the part I most want to see exist.
That funding will come from company sponsorships and individual contributions through Open Collective, and there’s a Discord open to anyone working on these problems who wants to compare notes.