Tweag

Incubating the Haskell Foundation

26 March 2021 — by Richard Eisenberg, Tim Sears, , Mathieu Boespflug

For the three of us, the launch of the Haskell Foundation was one of 2020’s few bright spots.

In November 2020, Simon Peyton Jones announced the formation of the Haskell Foundation. Conceived during the first lockdowns of early 2020, the Haskell Foundation is a non-profit organization aimed at promoting adoption of the Haskell programming language and bolstering its core infrastructure. The Foundation has now raised approximately $500K in cash and financial commitments from sponsors, recruited a volunteer board of 14 directors and a full-time staff of two: Andrew Boardman, Executive Director (ED); and Emily Pillmore, Chief Technology Officer (CTO). Today, the Foundation has its own point of view, plans, sponsors, and community of volunteers. It has reached escape velocity. Our personal opinions now or back then don’t matter much, but we thought it might be nice to share the story of what sparked its inception.

In the end, the Haskell Foundation was launched with input from dozens of friends of Haskell. What we thought might be like rolling a boulder uphill turned out to be more like kicking a stone and starting a rockslide. It seems the Haskell Foundation is an idea whose time had come.

Inception

In early 2020, Simon Peyton Jones got a call from Frank Rodorigo, a US-based entrepreneur, who was in the process of reviving Skills Matter, a community of tech creators, users and adopters. Skills Matter had run into financial difficulties in 2019, and Frank, together with his CTO Scott Conley, wanted to make sure Haskell was at the center of their reboot plans. He hoped to explore some ideas with Simon about the best ways to help Haskell. One thing led to the next: under the impetus of the rebooted Skills Matter, we brainstormed about what extra glue the community might need to bolster the lofty goals that so many seemed to have.

Over the years, we have been encouraged by the inspiration Haskell has given to other languages like Rust, Apple’s Swift and others. We even saw Java programmers singing the praises of lazy streams and anonymous functions. It had always seemed like just a matter of time before more programmers start wanting to use “The Real Thing”. While we saw some progress in the Haskell ecosystem, with the release of GHC 8.0, and efforts to eliminate “dependency hell” on the part of the Stack and Cabal projects, it didn’t feel like enough.

Enough for what? If you wanted to use Haskell in production at a company, you still had to be brave and determined. The evidence was anecdotal but hard to ignore. It ranged from conference attendees talking about how complicated Haskell has become, to concerns such as getting Haddock working with GHC’s new type system features, to the trouble enabling Stack to keep working with Hackage. Pull requests to some of the core libraries were languishing and many felt there were problems in those libraries simply going unaddressed.

Worst of all, some of us were aware of companies that had adopted and then later abandoned Haskell. Those would-be Haskellers faced a confusing collection of projects and committees, none of whom themselves felt they had a broad mandate to advance Haskell.

To be sure, there were some people thinking about remedies with proposals like Simple Haskell and Boring Haskell. But we thought more was needed. We started with some goals:

  • An easier on-ramp. Starting out with Haskell is harder than it should be, with a wealth of ways of setting up a Haskell dev machine (some of them out-of-date!). Even after this first step, newcomers often land on unmaintained wiki pages or other seldom helpful destinations.

  • More inclusivity. To help the community grow, we wanted the occasional alienating post or dismissive comment to be reliably addressed.

  • More progress, faster. We wanted to encourage more innovation across the many different projects that comprise Haskell. We also wanted to explore ways to help the committees and volunteers that make up the Haskell community to channel resources and volunteers to where they are needed most, and to ensure that each tool works well with others, forming a cohesive whole.

  • Funding. We knew first-hand that many companies were already investing in Haskell to ease their own pain points, but their efforts weren’t very connected. We wanted to make all of the above more feasible with the help of sponsors.

Gaining momentum

Having envisioned the outlines of the Haskell Foundation, what next? We wrote those ideas down in a live shared document we called “the whitepaper” and started gathering feedback from an ever widening group. We really wanted to iron out the wrinkles to avoid announcing something that could fall flat on its face.

As Simon Peyton Jones is the central architect and developer behind the Haskell ecosystem, the rest of us thought that his visible involvement and leadership would be essential. Happily, Simon agreed to stay involved and eventually spent way more time than he had planned.

We still had some worries. Would others consider affiliating? Were we stepping on anyone’s toes? We decided to take a slow and deliberate concentric rings approach to next reach out to core committees, then companies and large projects, then influential community members, to socialize the idea further.

We started with the Haskell.Org Committee. Expecting pushback, we were blown away by how eager the chairperson (Jasper van der Jeugt) was, and then in turn the rest of the haskell.org members: Ryan Trinkle, Emily Pillmore, Tikhon Jelvis, Rebecca Skinner, and Alex Garcia de Oliveira. They all became early and essential participants in the launch of the Foundation. It seems that the idea resonated strongly with some of their own ongoing discussions.

We were still a small team at this point — fewer than ten people, with Tim Sears as our chief day-to-day organizer. We kept adding anyone who wanted to help to a bi-weekly video call and kept iterating the whitepaper. After a short time it looked nothing like the first draft.

Feeling a bit braver, we then decided to also reach out to companies and key stakeholders in the community. We met skeptics along the way, like Michael Snoyman (now on the Foundation Board), both on the vision and the feasibility. These skeptics’ input turned out to be extremely helpful. Among other things, we used their feedback to sharpen the Foundation’s commitment to transparency. Ryan Trinkle (Obsidian Systems) soon started playing the role of shadow CFO. Duncan Coutts and Ben Gamari (Well-Typed) provided valuable input, as did Neil Mitchell, John Wiegley, Ed Kmett, Simon Marlow and others too numerous to mention.

It was important to connect with the Core Libraries Committee, the Hackage Trustees and the GHC Steering Committee. Emissaries were dispatched. Those groups ranged from amenable to enthusiastic, but they also asked some thorny questions. Did the community really need another committee? How would the Foundation differentiate itself? Could we actually raise money? In private, Simon Peyton Jones asked Tim why he wrote down a 7-figure sponsorship goal in an early draft of the whitepaper. Was it realistic? Tim had to admit he wasn’t sure, but without funding the Foundation would never have big impact. Only one way to find out…

At some point we started calling our informal cabal the HF Working Group. Eventually the invite list numbered in the dozens, with about 12 or so turning up regularly to our video chats. Scheduling was rarely a problem, since nobody was traveling - the silver lining of a terrible pandemic.

The Foundation escapes quarantine

In August, the Working Group asked itself: “Haven’t we socialized this idea enough yet? Can’t we freeze and ratify the whitepaper? Why can’t we just launch?“. Just like that we turned a corner. We started a semi-public outreach effort on the following basis:

  • The Foundation would be non-profit. Our goal would never be to create a consultancy or training company. Our goal would be to promote Haskell and related technologies. We would not be selling any services.
  • The Foundation would be inclusive. It would seek input from a variety of sources and be community-driven. A goal of the Foundation would be to look like the community we want to become.
  • The Foundation would be funded. We would start with ambitious fund-raising goals. Specifically, we wanted to aim for a yearly budget of over $1,000,000. Donations would come from industry and from the general Haskelling public.
  • The Foundation would have an executive director. Acknowledging that we are all busy people, well-meaning volunteers simply do not have the bandwidth to offer sustained attention to where it is needed. Instead, the Foundation would have at least one full-time employee, whose day job it is to manage the Haskell community and promote its interests.

To our eyes, now informed by the community, this seemed like a winning formula — a design that would be able to fix Haskell’s problems and promote the language, while strengthening our community.

The Foundation quickly became the world’s worst-kept secret as the Working Group set a public launch date for November.

Announcement and bootstrapping

Right away there was a new chicken-and-egg problem: we wanted the Board of Foundation to be drawn from the wider community, and yet we needed to have someone in charge so we could launch quickly. The HF Working Group landed on a two-step process. The Foundation would start with an interim Board and launch. The Board would then both replace itself and hire an ED to run the organization. The Working Group identified eight prominent Haskellers pre-launch and invited them to serve as an interim board. Thanks to Simon Peyton Jones’s persuasive powers, and a promise that their main duty would be limited to the above, they were quickly recruited.

Tim Sears, Emily Pillmore, Ryan Trinkle, and Alex Garcia de Oliveira led the fundraising efforts, starting seriously in August with a launch date slated for November. Even before launch, the Foundation quickly landed over $200,000 in commitments. Soon we knew that we would have enough to fund an ED who could spend enough time to foster the sponsorship efforts. Ryan started the paperwork to incorporate the Haskell Foundation as a non-profit, while Emily took over the de facto running of the Working Group. Jasper led an Affiliation Track to coordinate transparency policies for the community.

Richard Eisenberg worked with other volunteers (Ben Gamari of Well-Typed, Davean Scies of xkcd, Emily, Moritz Angermann of IOHK, Tikhon Jelvis from the Haskell.Org committee, and Tim Sears) to develop an initial technical agenda, a list of the kinds of projects the Foundation would seek to accelerate. This was to be used in our fundraising pitches and in forming a starting place for the real work to come: we knew that, once the Foundation was made public and we had selected a board, the agenda could be revised in the light of the freedom of being able to consult the public.

Rebecca Skinner from the Haskell.Org committee (helped by Tim and Davean) volunteered to lead up the effort to create a website, in advance of the upcoming launch, which we decided to incorporate into the Haskell eXchange conference, hosted by Tweag partner Skills Matter. Cardano/IOHK provided much of the on-the-ground labor in putting the initial website draft together, handing it over to Rebecca to push it over the line for the launch.

Finally, on 4 November, 2020 Simon Peyton Jones publicly announced the Haskell Foundation. You can watch the video of his announcement. Despite the feverish pace of work leading up to that announcement, everyone knew that the real work was only beginning, but we had a launch!

The story we wanted to share now comes to an end.

Epilogue

In December and January the interim Board recruited an outstanding slate of 14 members from a large group of applicants. As a special bonus they decided that the initial funding was sufficient to hire not only an ED, but also a CTO! Andrew Boardman joined as ED and Emily Pillmore stayed on in a permanent role as CTO. In its first meeting, the Board elected Richard Eisenberg as its chair.

The work of the Haskell Foundation proper is finally underway. This includes developing a process for community input, identifying projects to fund and otherwise support, and continuing outreach efforts. We expect that the creation of the Foundation will mark an inflection point in the history of Haskell. It is extremely gratifying to have played a role in helping a community that has given so much to us personally. We’re eager to see where it will go from here.

Finally, the Haskell Foundation needs volunteers, people just like you. The best place to reach the Foundation is via the Haskell Foundation category at the Haskell Discourse instance, though you can reach out directly to the Board or its chair, Andrew the ED, or Emily the CTO.

About the authors
Richard Eisenberg
Tim Sears
Mathieu BoespflugMathieu is the CEO and founder of Tweag.

If you enjoyed this article, you might be interested in joining the Tweag team.

This article is licensed under a Creative Commons Attribution 4.0 International license.

Company

AboutOpen SourceCareersContact Us

Connect with us

© 2024 Modus Create, LLC

Privacy PolicySitemap