Haskell Foundation Q1 2025 Update

Haskell Foundation
1

2024 Context

Financial Outlook

The end of 2024 was a challenging time for Open Source generally and the Haskell Foundation was no exception. Open Source foundations struggled to raise funds in the high interest-rate economic environment. LWN has a detailed article about some of the financial troubles faced by OS ecosystems here. The Haskell Foundation is not immune to these financial headwinds, but it is much smaller which is a strength in this particular instance.

The end of 2024 was a time of hard work within the HF, the Board and myself worked to raise sufficient funds in order to continue the Foundation’s work within the Haskell Community.

Toward this aim, we tried a few new things (new for the HF). First, we applied for a large NSF grant via the NSF POSE Program, we feel that the HF fits squarely within the scope of this program. If we do win this grant, it would be a significant improvement to the sustainability of the Haskell Ecosystem. Second, we reached out to many of the industrial users of Haskell and worked with them on how the Haskell Foundation could better meet their concerns. One of the outcomes of those discussions was the creation of the Haskell Foundation Advisory Board, more on that below.

Infrastructure

In 2024 we learned that our long time infrastructure partner, Equinix, was going to sunset its ‘Metal’ product. We had been a beneficiary of free credits for this product for many years, at no small cost to Equinix. Because of the wind-down of this product we were asked to migrate our services away from Equinix Metal in Q1+Q2 2025.

Q1 2025

NSF POSE

Awardees to the NSF POSE program were meant to be notified ‘early 2025’. However, due to the uncertainty in the US Federal government, all notifications have been paused. You can see the list of awardees here. While we remain optimistic about our chances, there is always the possibility that the program is cancelled outright.

Advisory Board

The Haskell Foundation held its first Advisory Board meeting, the result of this meeting was a very clear consensus that the Haskell Foundation should focus its time and energy on the smooth transition off of Equinix Metal and onto its new infrastructure.

Infrastructure

We are using the need to migrate our infrastructure as an opportunity to rethink our infrastructure strategy. As such the HF has worked with our infrastructure team and volunteers to ensure that all necessary services have a home and a plan for migrating off of Equinix Metal and onto a new platform.

This included a trip to Cape Cod in order to manually install a server for backups see photos here.

While the migration is not complete, I am happy to say that the plan is on track for the deadline from Equinix. There are a lot of moving parts, and no one person has all the keys to the kingdom, so coordination is the name of the game here. I can’t stress enough how great @bgamari, @chreekat, @davean, and @sclv have been in working through this progress. We all owe them a big thanks!

Conclusion

Overall I am optimistic about the state of the Haskell Foundation at the end of Q1 2025. There is still fundraising to do, and there is still uncertainty around the NSF funding, but I feel that we are up to the task.

Standard Update links and posts

Cabal-install 3.14.2.0 released

[ANN] Haskell Language Server 2.10.0.0

New projects

Existing initiatives

DevOps

Stability Working Group

Meeting notes:

Podcast

The Haskell Interlude Podcast released an interview with Conal Elliott and Farhad Mehta.

Sponsorship

All of our work is made possible by our individual contributors and our sponsors. Thank you so much for your support.

Gold

Silver

Bronze

33 Likes
2

Thanks for your work and for this informative update!

Please tell us more about the backup server in Cape Cod ? It sounds intriguing !

5 Likes
3

Apologies, I wasn’t trying to be vague.

Backups aren’t new for us, what’s new is that we are hosting backups (and some other services) at a co-location facility. This is more cost-effective for the Haskell community, but one of the costs is that we have to go there in person to do the install :smiley:

The location (Cape Cod) was chosen for a few reasons, but mostly a combination of price and the property that several of us are within 1 day’s drive.

5 Likes
4

Can’t help but noticing that the list of sponsors has shrinked significantly :frowning: Also, how did you decide to switch away from Monad/Applicative/Functor tiers to Gold/Silver/Bronze?

I hope that the need for migration will be used to set up more transparent processes in the core Haskell infrastructure. It has been frustrating to read threads like this one over the years, which show that the current setup is opaque and probably lacks that kind of automation that is considered baseline devops in the current century.

2 Likes
5

Yes, the list of sponsors has indeed shrunk, unfortunately. This is something we are working on improving.

Regarding the change from Applicative/Functor/Monad, many of the people who are decision makers within these companies don’t necessarily know what these things mean, whereas “Bronze/Silver/Gold” has a wider cultural base of understanding.

As for the point about transparency in our infrastructure. That is part of why the HF is working to support better coordination across Haskell’s various infrastructure groups. Many of these groups have been volunteering their time and energy for far longer than the HF has existed. We aren’t trying to tell them what to do, only to help ensure that they have the resources they need and to facilitate coordination/communication across teams.

8 Likes
6

The issue isn’t transparency, though we can work on that. The issue is that we do not have enough people to do the things we need – our team is skilled, but busy, and remains predominantly volunteer. We don’t want to be “opaque” – if you have any questions at all, ask!

But to improve we mainly don’t need better processes at this point. We need more resources! We need skilled volunteers willing to consistently (not drive-by) donate time and “own” subportions of our infrastructure. And we need money to pay for more non-volunteer time.

Otherwise, the transparent answer will all too often be “yep, that’s a good problem, thanks for pointing it out, we’re to busy to deal with it.”

4 Likes
7

Which subportions of the infrastructure need attention and what would be expected of new contributors?

Perhaps you could take inspiration from the CLC elections: CLC Elections January 2025. I found that very clear and easy to apply to.

1 Like
8

Perhaps it would also be good if there’d be a repo with infrastructure as code, that would also lower the barrier of entry. (Of course only if it already exists)

6 Likes
9

Thanks for the update! As someone newer to the community, I really appreciate the high level overview.

2 Likes
10

That’s why I spent a lot of extra time during the Stackage adoption creating a public IaC repository.

There does exist another such repo from before my time, but its security model is to only allow trusted access to root admins, so it can’t be shared. It is still the place where most Haskell services are defined.

Would it be valuable for me to work towards migrating more stuff to this public repo? One counterargument is that not many people actually do contribute voluntarily to infra, so it might be wasted effort. A) it’s hard B) it’s not shiny.

To be clear, this is not my argument. But hearing more opinions would be good.

6 Likes