Runrig

Appearance

Roadmap for 2024 ​

This roadmap is being presented in order to provide a more intentional, structured exposition of Runrig's short and long term goals. It is not, however, meant to be a static declaration of Runrig's purpose or motivations. It's a starting point, not just for the year 2024 but for each and every time we come together in community, and only for as long as we collectively choose not to plot another course.

Here's the TL;DR of what's covered:

Feel free to jump to the sections most relevant to your interests.

There will be a live-streamed & recorded presentation that covers this roadmap and what else lies ahead for Runrig in 2024 on Friday, Jan 5, 2pm Eastern Time. See the event page for calendar options and the call link, or find the live-stream and recording on the Runrig YouTube Channel.

I'm also seeking feedback, suggestions, and other comments on this and Runrig more generally, which can be posted on the GitHub Discussion for this this roadmap.

Lessons Learned in 2023 ​

Runrig got started in late-2022 first as a series of private journal entries and one-on-one conversations with friends, then as a private GitHub repo labeled draft-proposal, which for convenience sake I eventually made public. For me it was an exercise in articulating the kind of tools I wanted to build, the principles that would guide that work, and the overall culture of care I wished to surround it all. Ultimately, I was just trying to answer the question: What would I want an ag-tech worker's co-op to look like, based on my own experiences working in agriculture, food service and open source software?

For the past year, Runrig has remained in that same exploratory state. Designs have evolved, plans have changed, and by its very nature as a method of simple exploration, it has been anything but stagnant. Yet having never committed to a firm course of action, I struggle to see how its fundamental disposition has progressed at all beyond those early drafts and conversations from one year ago. All may be flux, but it still feels like stepping into the same river.

I've hesitated to center the initial concept and motivations for Runrig out of a concern that mine or anyone else's perspective might come to dominate and push out other deserving voices. While I wasn't exactly wrong in that impulse to make this an open and shared space, I was wrong about what it means to be truly open and sharing. I was reluctant to fully commit myself and take ownership of the project, because I knew how hard it can be to let go of that control as others come onboard and take a shared interest in its success. Avoiding all ownership and control, however, was no better a response than ignoring their perils entirely. The hardest thing, I realize now, is to accept accountability and be willing to commit to stewarding Runrig towards its next phase — call that "ownership," if you will — even while acknowledging that at some point I may have to let go of that control and either share or wholly cede its ownership to others. I hope when that time comes, I can do so willingly and joyfully, just like so many of my peers and antecedents have gifted me the dearly earned fruits of their own past labors.

With that in mind, I want to continue building those relationships, grounded in trust, which are so necessary to make this more than just a vanity project or wayward social experiment. And to do that, I intend to refocus my efforts towards developing the infrastructure and resources required to follow through on my commitments and solidify those relationships. While I still hope Runrig can provide a space for others and myself to play with new ideas and experiments, for at least the next year I've laid out a more concrete path of development, which follows below. Likewise, while I want our workshops to remain free and open to new projects and discussions, I'll strive to provide more intentionality, structure and focus to those events, which I believe will go a long way to making them more welcoming and inclusive to all.

I'm also making a conscious effort to use the first-person singular here. (Hi, it's Jamie! 👋🏻🦤). I just learned that the really hardcore grammatical term for that is clusivity, where "we" is inclusive and "I" is exclusive. I like to strive for inclusivity always, but this whole process — from a year ago to five minutes ago when I looked that up — has me realizing that sometimes the first step in bringing more people into unity can be to step forward and introduce yourself, as your totally unique self, before stepping back into your role within the larger group. Not always, but sometimes, like now. Plus, that whole "royal we" thing, grammatically inclusive or not, can come across pretty authoritarian and not at all welcoming. I know that's all a bit off-topic, but despite wishing otherwise and acknowledging just how many awesome people have contributed to this already, Runrig is still 90% just Jamie talking into the void, at least most weeks. And that's cool, but to make this a space where others not only feel welcome but have a sense of shared ownership, I think that takes a little more candor than I've been offering to date.

Three-Phase Business Plan ​

In the preliminary draft of "Appendix: Ecology" § "Main Legal Entities Constituting the Runrig System" in The Runrig Plan, I outlined 3 types of legal entities and 3 cooperative membership classes:

  1. A worker cooperative of engineers, designers and system administrators (aka, worker members), responsible for the building out and maintaining platform cooperatives' core services and end user applications.
  2. Platform cooperatives customized to meet the needs of each regional foodshed, providing a home platform for farmers and food workers (aka, service members) to access Runrig's network-based services and to host other FOSS cloud services.
  3. A data cooperative or data trust, whose members are users of the global data provider, which in turn serves as a centralized repository of all users data. These data members retain full control over how their own data is accessed, modified and shared, but in addition they enjoy governance rights over the management and administration of the provider's infrastructure, development, pricing, etc.

I've re-ordered them here to reflect the order in which they will most likely be incorporated. The worker cooperative, of which I myself hope to become a worker-member, would likely come first since it's needed to build the infrastructure for the rest. The latter two may develop in parallel, depending upon needs, but for now I think its safe to say individual platform cooperatives represent less complexity to implement, and so those I've ordered second with the data co-op or trust third.

It seems sensible that the Runrig worker cooperative, as the leading cusp of this triangular cooperative arrangement, would adopt a business plan to be implemented in three corresponding phases:

  • Phase One: Middleware, a.k.a. "glue services"
    • Targeted interventions that can catch the spillover of wontfix issues or otherwise unfulfilled feature requests from more mature FOSS projects, while also providing reusable middleware and extensions to connect previously incompatible APIs and services.
  • Phase Two: The One-stop Platform Co-op Shop™️
    • Technical assistance and consulting for the deployment and administration of platform cooperatives. This is analogous to the classic "web dev shop," particularly the diverse community of Drupal developers and consultants. These shops have proven that such an ecosystem can not only be sustained by free and open source software, even licensed under the "copyleft" GPL, but that it can positively thrive on account of the virtuous cycles that arise from the software commons and a precompetitive community of practice. The same potential exists for the semantic web and platform co-ops. To help foster its emergence, Runrig can provide a wide range of solutions: from fully managed hosting and technical support, to the hands-off delivery of complete systems that farms and food businesses can run for themselves; from crowdfunded feature enhancements and major upgrades for core libraries and services, to the same kind of small integrations of Phase One and special plugins that may only suit one client's needs; everything from the turnkey to the bespoke.
  • Phase Three: One Big Data Co-op 🐈‍⬛
    • Communally owned and controlled infrastructure for food sovereignty in the form of a cooperatively owned data cooperative or trust. This infrastructure is essentially what's termed the data provider in The Runrig Plan (see "Three Layers of Autonomy"). The most potential may lie in the form of a large Solid Pod Provider or other Linked Data Platform, but there may also be benefits to hosting FOSS alternatives to more conventional cloud services.[1] To be clear, complete ownership, governance, and control of these utilities will belong the co-op members — ie, the farmers, food workers, small businesses and end consumers who store their data there — not to Runrig. However, Runrig can be contracted to administer, maintain and further develop that infrastructure on behalf of its members. There may be other peripheral services required to deploy such infrastructure and optimize its connections to the platform services of Phase Two.

Preliminary to Phase One and the legal incorporation of the worker cooperative, it seems wise to delineate a Phase Zero, while we're still an entirely volunteer effort and prior to any financial transactions:

Of course, at such an early stage the reference implementations for the data provider may need to be scaled back considerably, or merely a proof-of-concept using an already available service, such as Inrupt.

The overall intention behind this three-phase plan is to provide a path for iterative development across the entire stack. The fact that Runrig's smallest, most rudimentary components are at the same time its most atomic and self-contained components is no accident: this is precisely what is meant by "Three Layers of Autonomy" in The Runrig Plan and why I think it offers unique advantages that cannot be achieved by your average VC-funded, proprietary tech startup.

Projects for 2024 ​

In the interest of getting Runrig past Phase Zero of the business plan, the main goal for 2024 is to build the reference implementations for the remaining phases. By no means is it intended to complete all of the implementations listed below; just one or two can be selected for each proposed phase, while the others can wait until funding is secured for that phase. I list them all here for context and to give a better sense of the potential for the years ahead.

Local Applications & Middleware ​

Reference implementations for Phase One:

Implementation of the Early Warning System is already well underway, as a part of the Crop Plan Auditing System we modelled for the Richland Gro-Op (RGO). The underlying Tauri app we've built could also be adapted to a general purpose system tray utility, that could provide notifications and link to third-party applications as a part of Phase Two.

Another concept we explored as a part of the Crop Plan Auditing System was the Dynamic Yield Coefficient, or DYC. It's fairly straight-forward, just a ratio expressing how many pounds of produce a given variety yields per bed-foot (or row-foot) planted. The challenge is in allowing the crop plan manager to adjust that ratio manually while ensuring that changes are propagated to all relevant computed data based on that ratio. In RGO's case, that propagation is achieved via spreadsheet cells and formulae, a fairly brittle system, but something we hoped to ameliorate by integrating her crop plan with a NocoDB instance or some other syncing mechanism.

Conceptually, this is all very closely related to an earlier NSF grant proposal that was never awarded for the Projection API, although I believe in our scenario it would be better termed the Projection Engine. This could essentially be a set of libraries for running constrained optimizations on crop plan data, something David Wright and I discussed might also be aided by his Rust port of ProcessScheduler. These utilities and the Projection Engine might not prove very useful in isolation, but could be incredibly powerful if we integrate with Qrop/Brinjel or Open Food Network in Phase Two.

Integrations with Third-party Apps & Services ​

Integrations with the following applications for Phase Two:

Brinjel (soon to supersede Qrop) is a desktop crop planning application sponsored by the French cooperative L'Atelier Paysan, and from the looks of it, they intend to provide cloud services once Brinjel is ready for general release. They're both licensed under the GPL/AGPL and are very well designed. Because Qrop is a desktop app running a simple SQLite database, it should be possible to integrate it with the Early Warning System.

Integration with farmOS and Field Kit, of course, should be fairly straight-forward. With farmOS.js we could stand up a really simple Node server to provide networked storage, or we can try adapting the farmOS Data Model to the Solid Protocol to provide network availability. Or both! Survey Stack can also provide some nice integrations without too much additional effort, since it also incorporates the farmOS.js client and modelling libraries.

I'm likewise very keen to seek integrations with Open Food Network, though this may not have any immediate advantages and I'm far less familiar with their API than I am with farmOS.

If a clear need arises, we can also try to connect the Early Warning System or other local applications to Airtable's REST API, or a NocoDB instance. In the latter case, we could just try connecting to NocoDB Cloud, their hosted service, or our own instance, if we get one deployed as a part of Phase Three.

Hosting & Data Services ​

Proofs-of-concept for Phase Three hosting solutions:

The potential of NocoDB is something that has appealed to me since a round of discussions with Tianna Kennedy of 607 CSA and the Catskills Agrarian Alliance. As a part of Skywoman's MAIA Project, we investigated a range of tech solutions but concluded that her team was already such a well-oiled machine that they really needed the full power of a spreadsheet application that they had full control over. Their one and only complaint was due to the brittleness of connecting many separate files each with potentially dozens of sheets, made all the more challenging when it came to importing new data from third-party applications.

One solution I proposed, but never really explored was what I called the "source-n-sink" spreadsheets or tables. It could work with spreadsheets, but seemed ideal to use with a more robust "no-code" database like Airtable, Notion or NocoDB. Here are my original notes:

3 part deliverable:

  • Script* for querying orders from ecommerce platforms (eg, Squarespace) and dump the raw data into a "Source Table" in Airtable. Immutable, append-only.
  • Airtable template(s) for "Sink Table(s)", which Tianna and others can use, change, manipulate, etc, but which draw on the raw "Source Tables" without changing them.
  • SOP for 607 CSA and other farms who want to replicate the process.

* = It's possible this script could be based on or integrated with Chris's eddie scripts.

I've come to think more that this kind of solution could also provide a sort of stop-gap to the kinds of challenges that Richland Gro-Op faces as well.

Ultimately, it's probably best to get as much data into more portable, semantic formats, like the kinds supported by the Solid Project. We could test on existing Solid pod providers like Inrupt, or stand up our own Community Solid Server (CSS) for testing.

Last but not least, all of these hosting scenarios could be greatly facilitated by collaborating with Co-op Cloud, who's lightyears ahead of us already in creating public interest infrastructure and what they call the configuration commons. They're really awesome people and tightly aligned with every single one of Runrig's core values. It would be to our great detriment not to seek a partnership with them and use their infrastructure wherever we can. They already have a NocoDB recipe and a farmOS recipe as a part of their configuration library, and we could surely contribute more as we find our way through Phase Three.

Building Community ​

It may be necessary to articulate a concrete program such as this in order to achieve Runrig's far-reaching goals, but I hope it will also invite new relationships, reinforce existing bonds, and strengthen the collaborative muscles required to make this a truly democratic and sustainable project.

Workshops & Mutual Aid ​

WARNING

The specific details of the workshop schedule and meeting links may be outdated but will not be updated. Please see the get involved page for the most current information about workshops and other events.

The Open Design Workshops so far have been held biweekly on weekend afternoons. While the consistency has been nice, I'd like to try alternating the days and times a bit to provide more varied opportunities for people coming from different timezones or who have a recurring obligation that might conflict with a single time and day of the week. We can always change it, but for now I'm scheduling it as 3 one-hour sessions each month:

Nth Day of the MonthOffice HrsWorkshop
First Friday7:30 - 8am ET8am - 9am ET
Second Sunday1:30 - 2pm ET2pm - 3pm ET
Third Thursday5:30 - 6pm ET6pm - 7pm ET

I am hoping the alliteration (plus how earlier days in the month correspond to earlier times of day) will make it slightly easier to remember, but of course I'll also keep a list of upcoming events posted here, as well as on the GOAT Community Calendar (via Google Calendar). I'm also trying out NixNix Events, a federated instance of Mobilizon, a social event planner made by Framasoft that supports the ActivityPub protocol. So you'll find all workshops posted in the Runrig Workshops Group.

The workshop format will be a little different as well. Instead of 2-hour free-form events, there will be a one hour workshop and/or livestream event with a set agenda, preceded by "office hours," where people can join half an hour early for more informal chat or to give-n-get a little Mutual Aid Tech Support (MATS). I'd like to see this kind of mutual aid work eventually have a greater share of the time, but it's an experiment and we'll just have to see how it plays out.

The workshop portion will be structured as a live-coding web stream, a report out on recent progress, or presentation on planned work for the future, depending on what projects we're working on and where they're at. In the beginning I suspect they'll be me talking and coding mostly, but as more people join we can shift to make it more participatory. I'll always try to leave time for introductions and check-ins/outs, and the main body of the agenda can always be adapted with group feedback in the time between one workshop and the next, based on what we're working on and who plans to attend. I'll try to post these to the Runrig YouTube Channel, too, so if people want to watch later they can.

Education, Training & Research ​

Along with the mutual aid work, I want to expand this to do events with more of a popular education component. I'd love Runrig to become a forum for regular skills-sharing sessions and technical training for those who have an interest in technology and who want to learn more, but also for community organizers and food producers to teach us techies and each other on the work their doing and how others can replicate democratic food distribution systems in their own communities. I believe it's also important to make space for political education, to really ground Runrig as a project for collective liberation, whether that takes the form of a book club movie watch parties, or some kind of trivia or game nights.

I've spoken with a few people about the idea of expanding The Runrig Plan into a robust design methodology, akin the Agile Manifesto or the Toyota Production System, but in conscious opposition to the capitalistic assumptions underpinning those philosophies. That kind of thing seems ripe for some form of Action Research (or PAR or CBPR). I've never done anything like that before personally, so it would have to be a truly self-organizing group effort, both for what we're trying to learn and how we're trying to learn it.

That will all take more than a year to come to fruition, but for now I think it's sufficient to look out for those opportunities when and where they arise. There are good reasons, after all, why "Education, Training, and Information" are together one of the International Cooperative Alliance's Seven Cooperative Principles:

Cooperatives provide education and training for their members, elected representatives, managers, and employees so they can contribute effectively to the development of their co-operatives. They inform the general public - particularly young people and opinion leaders - about the nature and benefits of co-operation.

Documentation & Communication ​

In a similar way, I hope within a year or two to be able to walk through Luis Razeto's How to Create a Solidarity Enterprise: A Theoretical-Practical Manual with my fellow worker-members in preparation for incorporating the Runrig workers' co-op. Given our skill-sets with design and web development, it could be an opportunity to create a more interactive tool for completing the exercises and business planning steps in the manual, which we could gift back to the community. How we publish our own documentation could share some of that infrastructure as well, presenting an opportunity to create new tools for garnering user feedback and for more members of the community to take an active role in the development of their technology.

Finally, with these other efforts to provide more structured events and documentation, it's important to maintain regular communication, so I hope to start sending out newsletters on a monthly or bi-monthly schedule with general updates on what's been going on and what's coming up. The newsletters will also be published to runrig.org, as yet another form of documentation, for folks who are dropping by for the first time, or for those who prefer not to subscribe.

  1. Examples may include Nextcloud for file sharing and storage (equivalent to Dropbox or Google Drive), NocoDB or Anytype as a low-code database/knowledge-base (Airtable, Notion, Obsidian), or Keycloak for identity and access management (Okta or Auth0). With time and concerted organizing, it may even become feasible to collectivize the kinds of cloud computing services dominated today by Amazon, Microsoft and Google, with FOSS alternatives like Ceph or Minio taking the place of Amazon S3 or Azure Blob as object storage services. After that there's the actual metal those services run on — the storage devices, networking hardware and server facilities — although I admit I'm getting way too far ahead of myself at that point. Importantly, however, even the smaller scale services like Nextcloud and Keycloak, while feasible to run on regional platform instances in Phase Two, will enjoy greater efficiencies by being managed on a global scale of the Phase Three data provider and provisioning resources out to the more ↩︎