Hi all,
the technical leadership committee of the Haskell Foundation has been successfully bootstrapped!
We have spent the last few weeks discussing our vision and strategy and are now fully operational.
We’re going to actively develop initiatives and proposals that we believe need funding and execution. For that we need companies, contractors and volunteers to help. Everyone can participate, whether it’s by dropping ideas informally, writing a concrete proposal themselves or finding entities that could support or fund initiatives.
We tried to keep the community proposal process as light as possible. However, we believe that fairness and participation are vital, especially when funding is involved, so anyone who writes a proposal should be aware that we will open the floor for the community to comment or even deliver counter-proposals. The process may evolve as we make experiences and get feedback.
Below is our vision. Feel free to get in touch with us!
The TLC’s shared vision
We believe that Haskell is awesome and that it can provide unique value to everyone, because it sits at an interesting intersection between human craftsmanship and machine automation:
- It is good for humans to write by hand: Type safety, powerful abstractions and referential transparency have enabled engineers all over the world in startups and big tech to deliver and maintain complex systems.
- It is good for responsible AI workflows: review is a significant bottleneck in modern AI-workflows and Haskell’s strengths heavily favor the reviewer. Additionally, the ability to quickly design eDSLs and express constraints where it matters allows to set up strong guardrails for code generation, without losing flexibility.
There may be better languages for formal verification or for vibecoding, but there’s no better language for “human in the loop”. Haskell is where we maintain control of our work, build institutional knowledge and ship responsibly.
We’re driven to improve Haskell for everyone:
- Beginners: the onboarding experience must be our benchmark for usability
- Experts: enabling advanced workflows and improving productivity
- Industry: focussing on stability and scaling
A vision of productivity
Imagine Haskell had the iteration speed of golang, the debugging capabilities of C# and the onboarding experience of rust. And all that while providing stronger guarantees and better performance!
Iteration speed
Modern workflows require fast iteration times: compilation, interactive repl, remote dependency caching, CI and test integration. All of this affects both human and machine workflows.
While Haskell scales well in terms of code complexity… it doesn’t scale well in terms of build times. As an example, GHC’s own CI has turnaround times of 6 hours. We want to bring them down to 2.
Debugging and profiling
Instead of helplessly dumping logs and binaries into LLM chats, we envision deterministic tools that allow precise and ergonomic enquiries:
- live step-through debugging
- actually useful stack traces
- beautiful performance analysis reports
And all of that in your editor/IDE as well as through powerful CLI and scripting interfaces, the latter enabling power users as well as LLM-driven debugging (natively, without requiring any custom adapters like MCP).
Debugging must be fun again, without littering Debug.Trace.trace all over your codebase.
Documentation and interactive exploration will assist newcomers in understanding all caveats of laziness and GHC and should clearly explain how to leverage these tools.
Onboarding and DevX
In these times of low frustration tolerance, we are going to build the best onboarding experience in the programming language ecosystem, because it is the ultimate benchmark for developer experience as a whole.
- installation so quick that you’re wondering whether the installer even worked
- documentation so comprehensive and accessible that you never get stuck
- tooling so intuitive that you don’t have to constantly ask for support
This can only be done if we commit to unix principles, which we summarize as:
- write programs that do one thing and do it well
- write programs that work together
- write programs that are extensible
Documentation should be a joy to read, written by humans, for humans.
A vision of a secure and trustworthy ecosystem
Security and trust don’t get established by accident. Once trust is lost, it’s often irrecoverably lost.
To protect our community assets and reputation, we need to:
- have the fastest incident response time among language ecosystems
- have penetration testers go through our infrastructure regularly
- make end users aware of vulnerabilities and mitigations
- build a robust cryptography library stack
- harden our supply chain
We should further support attempts to utilize or extend the strengths of the Haskell language to boost practical applications of secure design and architecture, such as Information Flow Control or Mandatory Access Control. We’re also monitoring the efforts of dependent types and liquidhaskell.
A vision of stability
What if we could compile Haskell software that’s 10 years old with the latest GHC, without any source code changes?
While this might seem like an impossible goal to achieve… embarking on that journey is going to get us closer to a robust, stable and predictable ecosystem anyway.
The questions we need to answer along the way are:
- how can users opt out of breaking changes?
- how long can we keep support windows open?
- how can we make the upgrade experience better?
Not only requires this investment in GHC and core tooling, but also potentially support community efforts to maintain stable package sets or coordinate ecosystem upgrades, make cabal more ergonomic for complex use cases, spearhead and support stability policies and gather data about what parts of the ecosystem cause churn.
We want Haskell to not only be a choice for fast moving startups, but also for large or conservative companies that do not want their workflows or software stacks disrupted. Stability will benefit both of these types of users, reduce cost and friction and build a reputation for Haskell as a modern language with a predictable ecosystem.
A vision of a flourishing open source ecosystem
All of the above is an attempt to support the Haskell open source ecosystem that we love. We are committed to open source, its values and the social aspect of it. Collaboration is how we engage with and inspire each other.
We want to enable all the maintainers and volunteers who keep this ecosystem alive, many of them unknown, doing their part on a Saturday evening at the other end of the world.
We code because we love to ship products. But we also code, because we want to socialize, meet people and have a good time. We’re committed to both of these values.