Proposing: Haskell Foundation Community Grants

Haskell Foundation
21

In case it matters…I’m not sure how much I really believe in the arguments I gave above…An approach that takes specific needs into consideration would be great for fostering goodwill, aside from the direct good it can do. I think my own suggested approach above is relatively cold and spartan.

People who want to do the work of organizing it would have my support to try any approach, if it’s just relatively small sums of money (at the start). If the energy is there, hopefully multiple programs can be tried out and tested in practice. I’d rather see multiple ideas tried that others have suggested here than none of them.

22

I think this is a great view of the overall situation. One thing I wonder about though is:

As a form of recognition to contributors. People contribute to open source all the time, for a variety of reasons. But the more ways we have to recognise, celebrate, and reward their work, the more likely we are to hit on the one that makes a difference for that person.

Would it be better to have the recognition piece separate from the grants? It seems to me that they are different enough that it would make sense to not try to meet both the goal of funding new work and the goal of recognizing past work with a single program.

2 Likes
23

This makes more and more sense to me, the more I think about it. The more that these grants are considered a prize for being a good contributor, the more we get:

  1. Discouragement, hurt feelings, and resentment.
  2. Pressure to spend donor money on something that doesn’t necessarily benefit the community because of fairness concerns.
  3. Loss of the opportunity for grants to bring new contributors into the ecosystem.

It would definitely be possible to set up a separate recognition program that doesn’t overlap with grants, and is significantly cheaper.

1 Like
24

I’m also a bit concerned about the “new contributors into the ecosystem” vs “help established projects develop larger features, etc” balance. I’ve been very much leaning to the latter. We have GSOC for the former, and the great benefit is not only that its their money, but also that its in part their outreach and the established brand. The most significant downside is that there’s a limitation to students.

I’m again not sure if money is the key thing for bringing in new contributors. It would be good to think through from scratch what the right things are – accessibility of projects is central, having projects organize hackathons, and tag good new contributor issues and features is an important component too. Also we have to ask if we’re targeting bringing in new contributors to haskell or existing haskellers into newly contributing to established community projects, etc. And doesn’t this overlap with that help-wanted index project that the HF is already undertaking as well? More questions than answers here, I confess…

4 Likes
25

I believe that personal grants have a huge potential to cause resentment and dissent. I also believe that having both HFTP and HFCG is likely to be confusing both for the community and for HF itself. My suggestion is to go with HFTP only, and after some time reassess how successful it is. We do not yet have an established good track record even for one form of community support; I think at the moment there are better chances to succeed without dispersing resources on many.

6 Likes
26

I share Bodigrim’s caution. We should be careful that our best intentions don’t become self-defeating. For example

manipulating people with incentives seems to work in the short run, it is a strategy that ultimately fails and even does lasting harm
…
people actually do inferior work when they are enticed with money, grades, or other incentives

Alfie Kohn - Punished By Rewards

Now, the book I’m quoting from isn’t exactly about the same situation as the HF finds itself in, but I think we should go in carefully and with our eyes open.

3 Likes
27

Couple more points.

  1. It might be nitpicking, but I do not quite agree with the wording. While HF Tech Proposals is by its name about tech improvements, HF Community Grants sounds like its foremost purpose is money. Yes, it’s possible that a tech proposal involves some funding, but it’s not in focus, while a community grant is meaningless unless we talk money first.

    We are likely to end up debating how to split a whooping sum of 500 USD between contributors, reviewers and maintainers from all over the world with different fees and taxation. It also opens a can of worms for collision and pressure, when contributors by various means persuade to merge, so that they can get money sooner.

    These risks are of course present in Tech Proposals as well, but I believe that by taking money out of spotlight we can significantly alleviate them.

  2. There is an important difference between grants in academia or industry and potential grants from HF. When, for instance, Tweag offers a grant, no one is entitled to raise questions how Tweag spends its money. However, HF intends to represent the community and spends community’s donations. It raises stakes immensely, when your grant application is accepted or rejected by HF, because it implies that it was accepted or rejected in the name of community, and everybody is entitled to discuss such decisions.

    Again, this risk is present for Tech Proposals as well, but more structured and relatively heavyweight process provides more possibilities to mitigate. A lightweight process, as Community Grants are suggested to be, is much more vulnerable to be perceived as unfair and biased.

I guess my question is: what exactly is so wrong with HFTP, which has not even started yet, that we already look for an alternative?

6 Likes
28

Thanks for mentioning this. I think it’s an excellent point. I’m not very convinced that GSOC is enough, but I’m far more convinced that focusing on new contributors is the wrong focus for a program like this just because we would be bad at it. We wouldn’t necessarily have the outreach to find a lot of new contributors. So I also feel that, whether we decide that way or not, a program like this would end up supporting mostly existing contributors spending more of their time on established projects. Of course, it’s a sliding scale in the end, and individual judgement would have to ultimately prevail.

1 Like
29

I’ll see if I can answer this without being too incendiary. :confused:

My goal in this proposal was precisely to restore the kind of support that the HFTP is likely to cost the Haskell Foundation, while also doing some good for projects in the process. From my outsider point of view, the HFTP looks positioned to debate and decide how the Haskell Foundation should spend its political capital trying to herd cats and get everyone onto the same page for important goals that are needed for the community, and worth some loss of good will to accomplish. Hopefully, that loss of good will can be minimized, but I’m not optimistic enough to deny that it will happen.

As you say, the HFTP isn’t really about providing support for people to do things. It’s apparently about deciding what needs to be done, and mobilizing people to do it, in cases where it’s not what the people involved were already motivated and active in doing. At least that’s what has happened up to now. Granted, sometimes it’s gone well (text-utf8), but other times it has been a disaster (unified installer) and bred hostility and resentment. And I’m not even entirely sure that wasn’t necessary. Some good things came out of the unified installer proposal, but it definitely came with a high price tag.

So the point here is, in some sense, to be the anti-HFTP. It’s to also let the Haskell foundation really step back and support the community in doing what it already does best, so that the community can do more of it. It’s positive reinforcement, to go along with the… okay, not really negative, but telling people to do something differently isn’t entirely positive either.

Another way to say this is: I’d like to be able to say “If you want to financially support the Haskell community, a good way of doing that is to donate to the Haskell Foundation.” Today, I have trouble saying that without a lot of qualifications. That’s because money donated to the Haskell Foundation doesn’t really do anything. As far as I can tell, it sits in a bank account, except for paying Andrew and Emily (who are being paid anyway from corporate donations) and $1000 of it that supported the HLS that one time. That’s over $2000 / month of my own donations that sit in that bank account, doing… I’m not sure what. I want to have faith that this is the right thing, and will help the community, but I’m making this proposal because at some point, I want that faith to pay off.

Sorry for being a little more forceful than I perhaps should have been. But I am afraid that perhaps the urgency of the situation is being lost. The HF is coming up on its one-year anniversary, and it still isn’t feeling as much like the community’s Foundation as I think it was originally represented.

4 Likes
30

I tend to agree with your overall set of thoughts and concerns, but I’m a bit more sanguine on the timing here. The one-year anniversary is going to be of the announcement, no? Like there was a large span of time where nobody was picked yet for CTO or CEO (and I believe the even extended the timetable on that) – after which there was starting to get the ducks in a row for some further structure. We had one set of projects as very initial projects in which to work out mechanisms and governance and sort of “learn the ropes” and that’s just sort of coming to fruition, and we’ve only just nailed down and approved an HFTP process at all.

So I’m not necessarily worried that things are going in a wrong direction so much as that they’re just taking a fair amount of time, and honestly I’m pretty sanguine about that, in the sense that when you have so many people and systems to coordinate, well, things do take time.

(edit: I should add that paying a CTO to run around and help coordinate discussions and organize projects and direct resources etc full time instead of needing to also have a day job has been really important and useful, and already very worthwhile in terms of keeping everything working, and paid off handsomely. But also, yes, covering core salaries isn’t a great incentive to encourage people to make individual donations, and I strongly agree that earmarked donations will be a much better incentive to make us individual donors feel our money is going somewhere concrete – personally I, like you, have been donating in kind to haskell for many years, and in the past i’ve donated to haskell dot org, etc. but as far as HF I’ve been holding off until there’s a particular concrete initiative I can get excited about).

2 Likes
31

@cdsmith thanks, much appreciated. I actually happen to share your concerns about HF in general, but not quite with regards to HFTP.

The HF is coming up on its one-year anniversary,

That’s not quite so, we barely passed half a year since February, when Andrew and Emily were hired.

As you say, the HFTP isn’t really about providing support for people to do things.

I understand your meaning here, but to avoid inadverent future misrepresentation I must clarify for other participants that this is not exactly what I say. I believe that HFTP is primary to support things to happen, not people by paying money (which is secondary or tertiary).

From my outsider point of view, the HFTP looks positioned to debate and decide how the Haskell Foundation should spend its political capital trying to herd cats and get everyone onto the same page for important goals that are needed for the community, and worth some loss of good will to accomplish.

This is not how I imagine HFTP and not how HFTP describes itself: it explicitly requires a consensus between parties. I think HFTP is actually very much what you describe as “anti-HFTP” :slight_smile:

2 Likes
32

I should add in my dream world the HF will employ not only a few execs, but also have a few part or full time engineers/sysadmins etc, and it could direct them to undertake projects (either smaller stuff like helping fixup the ecosystem for a ghc release, or individual smaller requested features to libs, or bigger stuff like significant improvements to tooling or libs). And the HFTP would have some “teeth” in that the HF could direct some of those resources directly to proposals.

3 Likes
33

Okay, as a thought exercise, perhaps what’s called for here is just a concerted effort to let people know (if indeed this is true) that it’s okay to write a modest proposal as part of HFTP, asking for support (financial or otherwise, I suppose) to do work on a Haskell project. Then I have the following concerns:

  1. The HFTP document itself doesn’t seem to be welcoming to this sort of thing. It says things like “Typically, this includes many rounds of discussion with HF leaders, Haskell library maintainers, and the broader community, several iterations on the design of the proposal, and some effort at prototyping the proposed change.” Now, perhaps the document is just incorrect about what is typical, should lots of modest proposals arrive. But it isn’t encouraging.
  2. While open and transparent discussion is something to be encouraged for working out a technical direction, it’s perhaps a bad choice for discussing someone’s need for child care in order to better contribute changes to a Haskell project.

The solution to these weaknesses seems to be in the HFTP process document itself, which offers this example: “An open source project requires funding for volunteers engaged in Summer of Code-type work on core tooling for Haskell”. Under this guidance, it seems that it’s most in the spirit of the HFTP process that this be an umbrella proposal for the entire grant program, which would be heavily discussed and then operated after accepted. But that gets us back where we started.

34

I have sent https://github.com/haskellfoundation/tech-proposals/pull/14 to propose this via the HFTP process. If you have further comments on the specific proposal I’ve written up there, please share them on that thread.

Thanks, everyone, for all of the input, both in this thread and in-person and private conversations. It’s been helpful in considering all of the options and making choices that I hope are the right ones.

4 Likes