I feel like haskell is dying

#84

I don’t think joining the haskell foundation will help much, though it would be nice if that was enough (joining on its own doesn’t resolve the issues which have led to the fractured community experience and ecosystem).

IMHO, I would summarize leading causes of attrition as:

  • Poor onboarding experience, primarily due to the community’s lack of cohesion and focus towards a successful onboarding experience, in-fighting, lack of clarity around goals, requirements, and who’s responsible for what, but also due to the stack/cabal schism.
  • Fallout related to the stack/stackage and cabal schism. This is a major contributor to a lot of problems. Most of us seem to have hoped/wished for it to blow over rather than actually be resolved. Now that it’s “blown over”, we can see that resolution is a better course of action.
  • The decentralized nature of haskell’s leadership (which is a really good thing in many respects), which we haven’t taken into account enough and the result contributes to attrition.
  • Frustrations getting agreement on anything significantly complex or controversial.
  • Lack of cohesiveness in goals and requirements related to defining GHC’s/base/haskell’s future.
  • We don’t care enough about the causes of burnout and attrition, and much of what leads to the burnout rate has been the norm for more than 10 years (or said another way, various “motiviated people” have pushed really hard for change over many years at a time and we didn’t care enough until they were so “done” they said “I don’t care anymore”).
6 Likes
#85

While it’s important to maintain perspective, “they have it worse” is not very helpful for our own self-reflection and understanding of what we need to do about the problem of attrition and contributor burnout.

3 Likes
#86

Hmm:

[…] presumably they [the HF] would also be interested in an accurate assessment/evaluation of the current circumstances […]

…@ketzacotal, maybe you can also help the HF on this matter.

1 Like
#87

What sort of resolution do you have in mind?

#88

That might be a topic for another (long) thread…

#89

LOL, yea, probably!

It’s not a topic I want to get more flames for, so I’ll keep my reply as simple as this for now:

There wouldn’t be a top-down schism, we would have a cohesive onboarding story with matching documentation on haskell.org, and onboarding newbs would know what to do, being aware of the options available to them, and feeling confident in the basics of how to build haskell software.

1 Like
#90

Well, you quoted only half of my post.

My point was that there’s nothing inherently unfixable in our community and we’re not at the stage where the only solution is to give up and regroup. Very far from it.

Additionally, I think discussing topics like stack-cabal schism don’t have any value anymore beyond bringing back some bad memories.

Onboarding experience has immensely improved, IMO. Tooling is finally getting better with more options (haskell.nix). There are a couple of people (e.g. @Kleidukos) driving documentation improvements. haskell.org page (esp. download page) has seen a lot of changes recently.

We’ve also discussed issues with GHC release management, HF has started an initiative about ecosystem stability concerns, etc. etc.

It’s not that any of these things are not recognized.

I think the main question is how can we create more engagement from volunteers and give those volunteers the space and help to drive change.

9 Likes
#91

At a risk of being a bit too dismissive, why would not you start contributing and make an informed opinion yourself?

Please stop scaremongering, there are no signs of attrition: plenty of core libraries continue to evolve and even more rapidly than before.

7 Likes
#92

I like how this thread of my rant exploded (Well, might be me making too much replies).

Sorry if my rants caused you be stressed or embarrassed, bemused. I wanted to voice my concern, but perhaps it has gone overboard.

EDIT: Oh no I sense some turbulence :cry:

#93

Why would you assume I am not speaking about legitimate issues taken from my own experience and those experiences from others I’ve witnessed or felt resonated with my own?

And I guess I should read your comment as implying we’re not welcome to share from our own experiences or contribute in ways that individuals feel are meaningful, even if they might dissent from your experience. Perhaps this perspective you are putting a voice to contributes to the problem?

LOL. No signs of attrition at all? I guess we must be crazy for even raising the topic for discussion.

One of the problems with this type of attrition (which is of newcomers more than seasoned experts) is that you don’t actually hear about the problem or why the newcomer decided to give up and move on to put their time and energy elsewhere.

1 Like
#94

There’s no shame in putting a voice to your concern and raising the topics for discussion. We should be thankful you were willing to put yourself out there in such a way, and we should use the opportunity to improve our experience.

1 Like
#95

I stand by my words: no signs of attrition, all major components evolve faster than earlier with a steady influx of newcomers. E. g., the majority of changes to bytestring last year were made by first-time contributors and not by long-term maintainers.

7 Likes
#96

There may be a simple explanation for these two opposing observations - Haskell’s net attrition is approximately zero, hence the relatively-flat graphs which @rae obtained for us. As @ketzacoatl has noted elsewhere, that could account for the departure of highly-experienced people from the Haskell ecosystem, while also accounting for the ongoing rate of activity e.g. on core libraries, as noted here by @Bodigrim.

Haskell was never intended to be “everything for everyone”, otherwise a more suitable name would have been UNCOL! As for “issues of management”…I refer people (again) to @rae’s earlier remarks and A History of Haskell.

…similarly, a five-year-old can be excused for merely complaining: unless they’re a child genius/prodigy, they shouldn’t really need detailed knowledge of “how things work”. As for the rest of us, we either have or usually can find that knowledge relatively quickly, so we should be able do more than just complain.

To those of you who have previous or current grievances with the current “management processes” : in addition to retelling your unfortunate experience, at the very least try to also be constructive by e.g:

  • telling us what you would have done differently if you were part of “the management” (or the manager),
  • or suggesting how (not just what) parts of the process can be improved.

This way, when new people interested in Haskell visits this or future threads, here or elsewhere, they aren’t greeted with this incessant “ringing of alarm bells”, then decide to look elsewhere - they can see us also trying to solve our problems.

8 Likes
#97

It really has, and I want to say a big public thank you to @hasufell because his project ghcup has been one of the main drivers of this improved onboarding experience!

9 Likes
#98

Dude, chill, Haskell is thriving, the sky is not about to fall. You don’t need to worry about stress or embarrassment on behalf of other people. We’ll all take care of ourselves. Take a deep breath and find some time to take care of yourself. Don’t worry: Haskell will continue to thrive even if you stop worrying about it for five minutes.

7 Likes
#99

Please do not do this. I want to see what people axually think and feel. If you create barriers for people that have negative thoughts and feelings (which is what you are doing right now), you create a bias in the information presented to the future visitor. You create a lie. This is not a way to care about the future visitor — this is a way to present a skewed picture of reality to further your own ends.

1 Like
#100

According to the sum of evidence, Haskell is not thriving. The case is not made. If you want to see a thriving programming language, check out Rust or Python. The graphs are starkly different.

2 Likes
#101

I refer you to the second message in this thread:

In any case, a hummingbird thrives and a blue whale thrives, despite the differences in their size.

5 Likes
#102

If you consider “thriving” as “can compete with rust and python”, then the answer is clearly: no.

If you look at the graph for the last 20 years, there’s no indication of a dying language.

3 Likes
#103

I’m curious to hear what “thriving” means to people and what Haskell would look like that would make you not worried about its future. Choosing Haskell or any other FP/FP-adjacent language is choosing something that is not available to the average programmer. Learning Haskell includes learning the effective use of a compiler, higher-kinded types, monads - to name but a few. Our part of the industry clearly has a market cap. I am not worried by that personally: I am a software engineer and I like code and abstraction, which Haskell is great for.

For example, the cooking industry has many different types of restaurants. You have many chefs working at low-barrier simpler skill cooking positions at chains. You also have struggling chefs working at specialty bakeries whose passion it is to bake the best cake. There is no “better” between the two - sometimes people want a burger, sometimes people want something new or unique.

Haskell is clearly the latter. We should try not to compare ourselves to other languages and talk about whether the Haskell community is doing as well as it can, not whether it is as good as the mainstream. I think @tomjaguarpaw is arguing that we are fulfilling our market cap quite reasonably. I generally agree, but if we want to talk about these problems, let’s accept these facts and talk about what we can do; not envy larger communities…

6 Likes