In the previous post I talked about what the Haskell Foundation has been up to for the first seven months, now I want to discuss where we’re heading.

The Haskell Foundation could undertake a vast array of initiatives, so restricting the scope of what we will use our resources on has been a major part of our recent efforts.

Mission

Amplify Haskell’s impact on humanity.

We have selected this mission statement to focus our strategy. It is general enough to make sure we can support the community in the ways we need to, but gives us something to measure proposed strategies against.

Strategic Focus: 2022

Increase the productivity of junior, professional Haskell developers.

For our first major strategy we looked to generate positive feedback loops. Not only do we want to amplify Haskell’s impact on humanity, but we want to improve our community’s ability to make changes that do the same.

This focus translates into a bias for accepting Haskell Foundation Tech Proposals (HFTPs) that relate to tooling enhancements that make junior, professional Haskellers, more productive. We will also create HFTPs as we find compelling use cases.

Update: To clarify, we mean developers who are junior in terms of professional Haskell usage, not in terms of their development career in general.

That does not mean that HFTPs need to be specific to junior, professional Haskellers. We chose this focus because we believe it will generate a broad array of additional improvements that benefit the entire community, while also ensuring that we get a specific set of improvements fully finished. Therefore, as we review proposals, we will prioritize those whose impact on junior, professional Haskellers is clear, but having a broader impact will be a plus.

Emily Pillmore and the Haskell Foundation Technical Track (HFTT) are responsible for both evaluating proposals from the community, as well as incubating ones where appropriate.

Tooling

I just read the phrase “algorithmic attention arms race” by Mark Xu Neyer, and it sums up the state of our world in a really elegant way. The best way to break out of the shackles of consuming culture is to make, and the most leveraged way to make is to produce better tools, to make making faster, easier, and more efficient.

If you caught my Haskell Love talk, you know that I feel passionately about developer tools, the developer experience, and the impact that has on our ability to make an impact in the larger world. To recap the ideas I presented there:

• Polished, functional, professional tooling allows developers to work at a higher abstraction level; the details they would otherwise have to juggle in their brains can be trivially accessible in their IDE.
• A tricky part of learning Haskell is understanding how the concepts fit together, and how they translate into the Haskell run time. An IDE that simulates the runtime and shows developers how their code translates into a running system not only helps programmers be more productive, it helps them learn the language better and faster.
• Making the language (and runtime) easier and faster to learn, understand, and debug, addresses some of the top reasons why Haskell’s reputation for production work can be rough.
• A truly interactive, simulating IDE experience also takes care of an issue for bigger projects: the edit -> compile -> run -> test loop needs to be instantaneous from the developer’s perspective for maximum productivity. We must not allow our tools to interrupt developer flow.

The Haskell language deserves an integrated development environment that takes advantage of its pure, lazy, strongly typed, functional programming fundamentals and provides an experience that delivers an order of magnitude improvement in productivity.

Junior, Professional Haskellers

We believe the sweet spot for focusing on tooling improvements is the workflow of professional Haskellers who are at the beginner to intermediate level.

• Hiring is both a strength and a weakness of our community. The self-selection bias of learning and sticking with Haskell gives a rich talent pool, but smaller than many languages.
• Senior Haskellers are rare, and those with experience translating business needs into production quality code rarer still.
• Managers therefore have their effectiveness gated on their ability to get quality work out of the beginner to intermediate engineers on their team, including their ability to hire them.
• Shortening the time for Haskellers to become seniors and leads both fills the existing talent gap, as well as makes our community as a whole much stronger.

Next Steps

The exciting new phase of Haskell Foundation operations is to identify, in our various task forces and committees, work that we can do to support our strategy, what resources to allocate to them, and get to work!

If you have an idea for a project that fits, are working on one already, or want to get involved, you can email the HF and we will direct you to the appropriate person or task force!

Seven Months!

It is hard to believe Emily Pillmore and I have been running the Haskell Foundation for seven months already. Similar to parenting, it feels like no time has elapsed, but at the same time it went very slowly.

We want to improve and become more effective, so in this monthly update let’s dive into what we’ve done over the last seven months, what we’ve learned, and where we want to go.

Fundraising

Our first challenge was raising funds so we could continue to have a Haskell Foundation, and it certainly took a while before the first check came in under our watch.

Last year, prior to the selection of the Board or the Executive Team, GitHub was the first check to clear, they had stepped up to continue funding GHC work that had previously been supported by Microsoft Research. IOHK came a month later with huge support for the Foundation, followed soon by Well-Typed, Mercury, Flipstone, Tweag, and Obsidian Systems. A month after, in January 2021, EMQ joined that illustrious group of early sponsors.

Fundraising has a long lag time between initial contact and checks clearing, so our next sponsors joined us in June, with Digital Asset joining us at the Monad level, and ExFreight at Applicative. This broke the drought, and we added TripShot in July, HERP in August (both as Functors), and CarbonCloud in September at Applicative.

Welcome to CarbonCloud as our newest Sponsor!

We talked to at least 37 companies at different stages of using Haskell, got tons of wonderful feedback and insight into what they’re doing, what is working, what is not. We converted those conversations into five new sponsors totaling $140,000, have another company that is in the final stage, and two more that are figuring out payment logistics.

Additionally, we were given an in-kind donation by MLabs: 40 hours / month of the amazing Koz Ross’s time to dedicate towards HF projects.

Lessons Learned

Fundraising is an ongoing process, and my primary focus as Executive Director. We are always looking for more companies to talk to and more opportunities to find funding for the Foundation and our initiatives. We are also now talking to existing sponsors about renewals and, where reasonable, increasing their contributions.

We cannot take our foot off the gas and relax, our resources are a fundamental limit to what we can accomplish.

Technical Track

Much of the work we need to do is fundamentally technical in nature, and we have largely been successful. We had some ideas of what we wanted to accomplish and did some deep dives into Backpack and the Windows platform for Haskell early on.

Utf8 Text

Andrew Lelechenko (aka Bodigrim) had a very focused proposal in mind: switch the internal Text representation from Utf-16 to Utf-8. This had been attempted before, but bikeshedding and arguing had stalled it out.

Bodigrim created his official proposal in May, disabled implicit fusion in Text in June (he found serious performance issues in basic cases while working on the changes), had a PR up for review in August, and merged the PR in September. Amazing and very well received work!

The next steps are PRs to the GHC codebase, and following the changes to downstream dependencies to ensure smooth updates.

Minimal Windows Installer

For a while Haskell support for Windows was a bit… rough. GHC put a lot of effort into fixing that up, but there was (and still is, but less so) spotty support by the tooling surrounding the compiler.

Julian Ospald stepped up to add proper Windows support for ghcup, and after a marathon of work and collaboration with Tamar Christina, Ben Gamari, and others, got it up and running!

If you want a maintained system GHC, Tamar’s Chocolatey package takes care of the complexity (but requires Chocolatey).

If you prefer to manage your own installation, and want a “system GHC” experience in Windows, you can now use ghcup.

For ease of use, managing multiple GHC installations seamlessly, or if you normally use Stack for your projects, Stack takes care of the complexity behind its CLI.

Lessons Learned

We attempted to use the momentum of this project to consolidate the ecosystem on a single Haskell installer, and unfortunately that did not go well. We did learn a lot from the experience, and made changes to how we go about selecting projects, getting community feedback, and how we assist with project management.

Haskell Foundation Tech Proposal Process

Emily Pillmore created a proposal process proposal, with a template to help guide members in the community on what needs to be thought out, decided, and written when proposing Haskell Foundation involvement in technical work.

It makes sure that we’ve thought through many of the issues that gave us problems: making sure the right people are notified, looking at prior art, determining motivation and deliverables, and what resources are needed to enable success.

Haskell Foundation Technical Task Force Elections

The HFTT is a volunteer group that evaluates the proposals. Emily posted a call for applications to join, received excellent results, and selected the new members. Participation in Haskell Foundation volunteer groups by a wide range of community members is crucial for making sure different points of view and perspectives are involved in the decision making, and we encourage everyone to get involved.

Extended Dependency Generation GHC Proposal

There are other proposal processes in our community as well, the most famous being the GHC Proposal Process. HF Board member Hécate Choutri found this gem, particularly given our love and support of the HLS project, and asked the HFTT to rally support for it. We absolutely agreed, and have requested that the GHC team prioritize it (within reason, given how slammed they are with getting releases out the door).

What does that mean? The Haskell Foundation provides support for GHC development (and we’d love to provide more, please donate and sponsor!), so we get some say in how the work is prioritized. We generally leave it to that team and the Steering Committee’s best judgement, but occasionally when we see a priority to help the ecosystem, we let them know what we’d like moved up in the queue.

Dashboard Proposal

Haskell is a very fast language, but occasionally a problem sneaks through CI and testing and ends up in production code. Emily worked with Ben Gamari to draft a proposal to take infrastructure the GHC team had in place, provide better UX, and extend it beyond GHC itself to cover important libraries that are a dependency of a large percentage of Haskell projects.

We want to consistently measure the performance of GHC itself, core libraries, fundamental dependencies, and eventually more. This will allow us to find GHC changes that affect library and application performance, address regressions closer to when the change is made, as well as lock in performance improvements. If you have expertise in DevOps, data visualization, and performance analysis, you can make a big difference here!

Cabal

A hole in our ecosystem had been consistent, inclusive maintainership of the Cabal projects. Emily stepped up to fill this need, did substantial work to update the code base to modern practices and styles, and found people to help maintain the project. This has led to new releases, new maintainers, and plans for future releases and features.

Core Libraries Committee

Similarly, the CLC had become operationally defunct, and once again Emily stepped in. She created a new way of working process, largely similar to the HFTP and the new process for Cabal, found which existing members were still active and able to perform their duties, and held an election process to find new members. Clear expectations have been set with a focus on communication and consistency.

Documentation

The fabulous Hécate Choutri does a fantastic job organizing volunteers around radically improving the documentation in our community. They have rallied efforts around the Haskell.org wiki, improvements in Haddocks, and Haskell School, a community sourced learning resource from the ground up. It has eight contributors, 76 commits (as of this writing), and three languages in development.

If you are passionate about bringing more people into our community faster, join in!

Performance Book

Gil Mizrahi’s Haskell Performance Tuning Book proposal has been submitted and is getting feedback. A critical gap in the knowledge of intermediate level Haskellers is the ability to deeply understand the performance of Haskell applications, how to address issues, how to design performant systems, and how to use the variety of tools available to debug issues.

This proposal offers a solution: a community sourced and maintained online book with the cumulative knowledge of Haskell performance experts, so that set of skills can be widely available and accessible. No matter what your level of expertise, you can help make this a reality!

Matchmaker

Matchmaker is project to address the issue of volunteers who have time and expertise but don’t know what projects need the help they can provide, and the maintainers who need that help and can’t find the volunteers. If you’re looking for a way to help out, here is a project that would help you help others in your exact situation!

Community

Haskell Interlude Podcast

Niki Vazou, Joachim Breitner, Andres Löh, Alejandro Serrano, and Wouter Swierstra proposed a long form podcast to interview guests about Haskell related topics. Originally they released a teaser episode, introducing the hosts, followed by Episode 1, an interview with Emily Pillmore. Episode 2 is out today, featuring Lennart Augustsson!

Code of Conduct

We started with the Guidelines for Respectful Communication as a foundation for how we wish interactions within the community to be conducted. We have determined that the GRC is necessary but not complete, and are working with the Rust Foundation to standardize on a Code of Conduct that would ideally be shared between the two communities.

This is meant to augment and enhance the GRC, not replace it. If you would like to be involved in making our community a friendlier, more inclusive place, please join our Slack and let us know in the #community-track channel.

Affiliations

Affiliation at this time involves adopting the GRC, but we’re also discussing HF endorsement of open source projects depending on the level of support the maintainers are willing to sign up for.

As an example, we could have multiple tiers:

* Core
* Security fixes turned around within 24 or 48 hours
* Stability guarantees, support for the last N releases of GHC
* Maintainership line of succession
* Code of Conduct / GRC adoption

Each level would include all the requirements below it, and would give people choosing which libraries and tools to adopt better information about the state of that project and whether they feel comfortable using it in production given their project’s needs.

Current Affiliated Projects and Teams

Haskell IDE Team
GHC Steering Committee
Clash Lang
Haskell Weekly
Core Libraries Committee
Haskell Love
Zurihac
Haskell.org Committee
IHP (Integrated Haskell Platform)
Stack
Stackage

Our Foundation Task Force

Chris Smith and Matthias Toepp have created a new task force aiming to help the community feel ownership of the Haskell Foundation. It is Our Foundation, not theirs (or mine). Their first initiatives are increasing the number of individuals donating to the Foundation, and a grant program to steer Foundation funds to the community. Please consider applying to be part of the task force!

Haskell Teachers’ Forum

A very early stage idea to bring together educators teaching Haskell at all levels. We want to share materials, best practices, and ideas. Send me an email if you’re interested in being involved!

State of Haskell Survey

Taylor Fausak has been the steward of the survey since 2017, and we’re discussing how to use HF resources to understand our community better. We would love more and better data about the state of our community, both so we can measure progress, as well as make more informed decisions.

Zurihac

On June 19th Emily gave a Haskell Foundation status update talk at Zurihac.

Haskell Love

My Haskell Love talk was on September 10th, where I talked about how our tooling is the linchpin for unlocking the potential of Haskell to solve real world problems and be the future of software engineering.

Office Hours

On the first Monday of each month, we host Haskell Foundation Office Hours at 16:00 UTC (9a US west coast) on Andrew’s Twitch channel. We have run two so far, the next one will be on October 4th. It is a ton of fun, we talk about a variety of topics, so please bring your questions and feedback!

Misc.

There are operational details that need to be taken care of, particularly true when we had just started. Benefits, payroll, legal forms, accounting, and so on. Big thanks to the HF Board Treasurer, Ryan Trinkle, for helping us navigate all of this and making sure we get paid, the HF gets paid, and we have all of the legalities settled.

Updates

• 2021–10–05: Added Stack and Stackage to the list of Affiliated projects.

Plus, that means people might read it.

Here we will feature interviews with Monad level sponsors and above, guest posts by members of our wonderful Haskell community, and updates about the HF and what we’re up to.