summary
TLDR: Open software and infrastructure are poorly supported within today’s academic environment. Meanwhile, market forces seem to drive almost every “open” company towards acquisition, enshittification, or both. As we learned more about worker cooperatives, and our individual careers fortuitously aligned, we decided to jump: we are starting a worker cooperative to build, maintain, teach, and support open source scientific imaging software.
Read on for more details, and our individual perspectives.
introduction
Each of us at Image Cooperative has been in an academic setting and seen cash-strapped universities shell out $100,000 or even $1M+ for licenses to proprietary software (e.g. Matlab), while letting talented developers go because they “can’t afford” their wages. Meanwhile, open alternatives like Python and NumPy get by on a dime and the unpaid work of thousands of volunteers.
There’s amazing science happening in academic and research institutions all over the world, and we are still driven to be part of it, and support it. But if not as employees at the institutions themselves, then how?
Over the past few years, some of us did some freelancing, which is lonely and precarious. We contemplated starting a conventional for-profit company, but we were apprehensive about the broken incentives in corporate structures: owners and workers are fundamentally at odds with each other. Unions are a powerful tool to balance those odds, but they often result in an adversarial relationship between workers and owners/CEOs.
Two years ago, we came across Valerie Young’s excellent talk, Inside Igalia: Scaling a Co-Op Beyond 100 Members. If you are still reading, you should probably add it to your watch list.
Although we had varying degrees of familiarity with worker cooperatives, we had not seriously considered one, and the talk provided a proof-of-existence for a long-lived coop in the open source tech consulting space. Valerie speculates about “federated Igalia” at the end of the talk, but it turns out they weren’t quite ready to act on that, or this post might be titled “why we started imagigalia.” 😅
Nonetheless, Valerie kindly volunteered her time and provided valuable advice on how we could get started. We then got in touch with the Cooperative Federation, an Australian peak body representing cooperatives in Australia, and itself a cooperative (a meta-cooperative, if you will!). We mention this because this is our first time doing something like this and their help was absolutely priceless. We do hope our story will be long enough to inspire others to create different cooperatives, and, if you are in Australia, please get in touch with them. Otherwise, the International Cooperative Alliance can help you find a local group near you.
We will write many more posts about how much we love cooperatives, but for now, let’s pivot back to what we aim to do with the cooperative.
We all came together over shared values of what management, analysis, infrastructure, and publication should look like for scientific image data. Individually, we’ve solved parts of these problems in small ways (working with individual labs on part of these elements) and big ways (collaborating on or leading projects such as scikit-image, OME-TIFF, OME-NGFF, and napari). Together, and with your help 😉, we hope to make it easy for all scientists, indeed all institutions, to use open and reproducible pipelines for image data, all the way from acquisition to publication and archiving.
our members’ perspectives
Coming from an academic background, we all feel a bit cringey about corporate communications that try to average everyone’s perspective and end up being a beige blur. So, we thought we’d include in this post some short individual, non-averaged thoughts from all five of our founding members.
Draga Doncila Pop
Over the past four years I’ve been lucky, privileged and honoured to be paid to work on Open Source Software (mainly, napari), while completing my PhD. Nearing the end of my PhD has been a time of reflection – on where I’ve come from, what I’ve been doing, and what I want to do next. On one hand, I love research, and everything it stands for. On the other, I’ve consistently found the academic system to be at odds with its purported goals of expanding humanity’s knowledge, and access to that knowledge.
I felt heavy pressure to publish work I didn’t believe was ready. I heard horror stories from fellow PhD students who were overworked and undervalued. I listened to other, more senior, members of the faculty lament on the state of academic publishing, the decline in the quality of teaching, and the constant battle for funding. My work on open source research software was consistently dismissed by my Advisory Panel — “it’s great that you’re doing this, but it’s not research.” Nevertheless, I contemplated an academic pathway, via a postdoc. Maybe I could struggle through the early years and build myself a niche where I could make meaningful contributions to knowledge and put those contributions in scientists’ hands via well-maintained, robust open software. But, deep down, it felt like taking the academic pathway would be tantamount to giving up.
In my other life, I’ve been serving on the napari Steering Council. I’ve seen first hand how difficult it is to get funding for Open Source Software – the funding opportunities are few, and projects that enable both academia and industry to succeed must fight for scraps. Time and time again when discussing funding, we questioned whether things would’ve been easier if we’d just decided to sell a product. But the life of napari, and other projects like it, is its community – we’re open, welcoming and always happy for others to join the fray. These are the qualities that make me love Open Source Software, and that made me love the collaborative work I did during my PhD. From a personal perspective though I had to ask – can I afford to commit to Open Source Software full time? Can I make a living from it, in the long term? Should I just “get a real job”? Another option that felt like giving up.
Image Coop doesn’t feel like giving up. It feels like an opportunity to keep doing the things I love, to keep enabling researchers and scientists to do the things they love, to work together to build something open, equitable and sustainable. I’m equal parts excited and terrified to start this journey. In my experience, that’s where the magic happens.
Eric Perlman
Throughout my career, I’ve worked on making image (and image-like) data open and accessible to everyone. While working at Janelia (where I met Juan) in Davi Bock’s lab, I worked on the imaging data pipeline that moved images from microscope, to storage, analysis, and researcher annotation with CATMAID, yielding the full adult fly brain (FAFB) dataset.
One critical, formative lesson for me is that value in making data public: FlyWire combined our published data with advances in image analysis, cutting edge web-based visualisation tools like Neuroglancer, and citizen science to produce the first full-brain Drosophila connectome. The response was immense, with media coverage including The New York Times, the BBC, The Gaurdian and even Saturday Night Live (seriously). For me, this concretely demonstrated the value of open data, open source software in science, and I’ve made it a part of my mission ever since.
Using my experience and connections from within this community, I was able to establish myself as a freelancer. This has given me the opportunity to work on many cool scientific problems with great people. It’s been a joy working across multiple groups all fundamentally driven by a love of science. (I’ve actually been encouraging Juan to try it for some time now!)
When Juan told me he was getting ready to actually take the leap, I was excited. Having grown up in Berkeley, I’m no stranger to workers’ cooperatives. In fact, my mom was a member of the now-legendary Cheese Board Collective. So having this organization be a cooperative is in some sense a return to my roots, which feels wonderful. And it’ll be good to combine the scientific freedom of freelancing with the collaborative spirit (and support) of a cooperative.
Kevin Yamauchi
Throughout my scientific career, I have been drawn to bioimaging. I love marveling at beautiful images of cells, tissues, and organisms and wondering how it all fits together. I have worked across many biology-adjacent fields ranging from biophysics to diagnostics, but the common thread that kept me excited was developing new methods and seeing them used to make observations and insights. From early on in my scientific career, I grappled with the latter: how do I share the methods I am developing outside of the lab I am working in?
When I discovered community-driven open source software, the light bulb turned on. Suddenly, I was able to collaborate with people from across the globe who were excited about the same things as me. No more wasting time reinventing the wheel. We pooled our different expertise to solve problems that none of us could have solved individually. Working on community-driven projects like napari and SpatialData really drove home for me how important it is that scientific software is owned by the community and not companies. Our continued scientific progress depends in part on our software evolving to meet the needs of cutting edge scientific questions and not the needs of shareholders.
As an academic, I felt a constant tension trying to balance the incentives that drive the academic system with the clear need for well-maintained open source scientific software. While I was fortunate to have mentors who supported my open source contributions, it was always clear it wasn’t the priority. In many conversations with my co-founders over the years, we wondered if there are ways we can organize the labor that supports open source scientific software that doesn’t feel like swimming upstream.
Co-founding the Image Cooperative feels like an exciting opportunity to try to establish a sustainable path for developers and maintainers of open source scientific software. I have no idea if this experiment will work, but I am feeling optimistic as I couldn’t ask for a kinder or more talented team of co-founders. I think it is clear that the way we write code is changing rapidly, so what better time to experiment with the systems that structure how we work?
Josh Moore
For the last 20 years I’ve been incredibly lucky to be supported to work full time on open source software. Almost all of that time has been spent on the OME project, which I think stands as a remarkable example of building production-quality software in the open, for the benefit of a global scientific community. Working alongside so many talented scientists, developers, and research software engineers has been without a doubt the great privilege of my career.
At the same time, my path through academia has always been a slightly unusual one. I never completed a PhD, and despite the impact of the work and the success of projects like OME, that has limited the formal roles available to me within academic institutions. More broadly, it has often felt like the academic system doesn’t quite know what to do with people whose primary contribution is building and maintaining research software.
Many years ago, while I was at the University of Dundee, representatives from funding bodies asked us whether we saw a clear career path for the work we were doing. At the time, we had to admit that we didn’t really know. In truth, even now, the long-term structures that support open scientific infrastructure still feel fragile.
The Image Cooperative is an opportunity to try something new. I’m not leaving my work with German BioImaging — which continues to be deeply meaningful to me and important to the community — but the cooperative offers a complementary path. For many senior research software engineers, especially those who care deeply about open infrastructure, it provides a model that could allow us to continue doing the work we believe in while building something sustainable.
Looking ahead, I would love to see the Image Cooperative grow, or even better, to see many more cooperatives emerge. My hope is that this becomes part of a broader ecosystem: communities of people building the tools they care about, supporting each other, and doing work they love for the right reasons.
Juan Nunez-Iglesias
I think a lot of my career can be described by the phrase, “there must be a better way!”, after which I would go down various rabbit holes, often to the detriment of my actual science questions, but, I hope, to the benefit of many others’. During my PhD, I gave a talk to my labmates about Subversion (the thing that was there before git, for the young’ins in the audience), and how we should stop having our code in folders named after the date, and for a couple of years the lab subversion server was my desktop. Later, I would sometimes spend months upstreaming code to open source libraries like scikit-image, to the chagrin of my PIs.
That kind of work really wasn’t valued in academic circles, resulting in precarious employment for most of my career. The only exception was a three year contract from Ian Harper at Monash Micro Imaging, for which I am forever grateful, and a fellowship from CZI, for which I am also forever grateful. The latter was renewed twice (more gratitude!), but as each renewal approached, I was told by Monash leadership that I wouldn’t have a job without it. Needless to say, that felt terrible, and I’d been contemplating freelancing since that first conversation, in 2021 — and with encouragement from and thanks to our co-founder Eric Perlman, who had been doing it for some years now.
But I had already done a bit of freelancing on the side in the past, and it was lonely, and felt even more precarious. I really wasn’t looking forward to it.
That’s when I came across Valerie’s talk (mentioned above). I watched it twice, breathlessly shared it on #randoms in the napari Zulip (open login required), and it’s been on my mind ever since, rent-free as they say. I shared it and discussed it as often as I could, including with this lot. 😂 By last August, I knew my fellowship was coming to an end and this was definitely do or die for me.
Many conversations later, here we are. Like Draga, I’m equal parts exhilarated and terrified. But, mostly I am, yet again, insanely grateful to have my co-founders with me and to be starting something we deeply believe in. “There must be a better way…” Is this it? I sure as hell hope so.
closing thoughts
We all believe in the ideals of the Academy, in the pursuit of knowledge for knowledge’s sake, and in the construction of things that improve the world and outlast us. There is an alternate timeline in which academics have stable jobs, are valued for their craft rather than their output, and none of us feels a need to go out and found this cooperative.
That is not our timeline. So we’re trying something else. We’re actually tremendously excited because, together, we have a really diverse set of skills to solve the problems we want to solve: data visualisation, data analysis, (bio)physical modelling, large data handling, cloud data, image data standards, open source development, and more.
So we have an excellent opportunity to push image analysis in science forward. If you’re struggling with image data, at any stage in the pipeline, please share this widely with your colleagues, and think about whether some of the proprietary license money you’re currently spending could be spent instead on open solutions that would be yours forever, endlessly adaptable, and benefit the entire world. And please get in touch by emailing [email protected].
If you think you want to work at Image Cooperative, please look at our joining page. We are at the very earliest stages of our history, and we need new contracts to grow, but our ambition is that, 20 years from now, we will also have a YouTube PeerTube talk: Inside Image Cooperative: how to scale a scientific imaging cooperative to over 100 members.