I feel like haskell is dying

Well put, I second that

2 Likes

Yea, companies might be naive. Just that… it is hard for me to ignore the companies. For one, I am not working in software industry as they would only accept ppl serving in the famous infrastructure. Scala is not on the table.

It does not help that the archetype of SW company in my region is a game company building NFT infrastructure. This company makes their own coins in a crude way, and proceeds to print more and sell them off for profit.

1 Like

Many languages include laziness. Map in Clojure is lazy, for example. F# also has laziness in the sequence type. Common Lisp can add laziness in a trivial manner.

Haskell is not a pure language - it just concentrates the impurity in small locations. That’s a good approach, irrespective of the language that you use.

1 Like

A good programmer can write Fortran in any language - or, in my case, writing VBA in the style of Common Lisp. :crazy_face:

I think this is a pretty complicated topic, but one that is worth exploring together.

While I agree there are plenty of indicators for Haskell’s positive growth (more haskellers, more use of haskell in industry, bigger FP impact on other languages, etc), there is a discrepancy between those indicators and the overall health of our ecosystem.

Rather than arguing whether or not haskell is dying, I think it would be much more useful to consider how healthy the haskell community and our ecosystem are, as well as what could be done to make it more resilient and more productive.

10 Likes

Well, a few key people left the community. Others of the old breed are still active in Haskell in industry, but stopped spending a lot of time on their open source life.

I think we’re experiencing a small generational shift, that’s all. New people have to pick stuff up and drive the train.

8 Likes

Is it really that simple? What if there were more to it?

This perspective comes across a bit too dismissive for me to agree that it really is ok.

For example, we should be looking closer at the what and why behind attrition (not just of the folks you speak of, but also in general).

What leads to attrition in the community of contributors?

Maybe some people would feel comfortable sharing their stories, or how they feel about being a contributor, or more precisely, why they started down that path and what knocked them off course?

3 Likes

You might be agreeing me on that one specific point, but I was talking perceptions, not realities.

I still think GHC ought to be better at making new features with!

I’m a huge believe in Things You Should Never Do, Part I – Joel on Software . And I think it applies to language implementations too. The consequences of this are profound, in that the older the implementation is, the more easy it should be to run experiments with new features, and also that new languages are just not going to be ready in ways that are hard to anticipate.

I think the only thing holding us back from this basic truth is a conservatism about what GHC is for what what it should like it. That’s it. Haskell is a great language for writing compilers, and being able to incremental evolve GHC with a its existing testsuite and the Haskell library ecosystem is a huge advantage!

4 Likes

I think I agree with both @hasufell and @ketzacoatl. On the one hand we will find that a lot of contributors stopped contributing because they grew up, had kids, starting running 50-person Haskell teams at FAANG, etc., and we can’t do anything about that. On the other hand it there will be some others who left for reasons that we can fix so it would be worth trying to investigate this.

8 Likes

Open source is a rough place. I’ve spent a couple of years of my free time in distro development. I can say whatever people think about problems the Haskell community has… there are way more toxic places.

Note that I’m not dismissing this point at all. It may be worth investigating this, but doing so in public often creates tension itself. This is something that can be gathered via surveys, etc. I’d certainly welcome a HF initiative of culture building. It’s probably the hardest task in tech, amongst people with strong opinions. And I think I expressed several times that I see HFs primary job in exactly that domain.

7 Likes

Hm. Last I checked, Clojure, F#, and Common Lisp are all becoming out of fashion - while stuffs like Go, Python and JS are getting more and more traction. Could be me though.

1 Like

Is this true? I want to believe.

Question - if Haskell did disappear, which one of these two groups would you rather be in:

  • the group which did nothing to prevent it,
  • or the group which did all they reasonably could to prevent it?

I’d be in the group which tried to prevent haskell from disappearing. For this task, though, it is important to accurately assess and evaluate the current circumstances.

1 Like

Then I suggest joining the Haskell Foundation - presumably they would also be interested in an accurate assessment/evaluation of the current circumstances, in order to e.g. determine where to direct resources.

3 Likes

Oh, how could I do that? I thought of it before, but it was perceived to me to have a high bar of entry.

From a quick web search:

You can also ask a member of the Haskell Foundation Working Group (e.g. @rae) for more information - just be patient; whoever you contact will be busy!

3 Likes

Thank you, I just sent an e-mail to the volunteer section.

1 Like

I agree that HF “membership” would be a good way of declaring one’s interest in promoting Haskell. But we don’t really have such a thing today! You can be a member of various committees or task forces, but those openings come up only occasionally and tend to be focused on a particular goal. I have created Implement a membership scheme · Issue #28 · haskellfoundation/tech-proposals · GitHub as a request to brainstorm a concrete design and implementation for membership.

5 Likes

…I only meant “member” as in you being someone involved in the HF Working Group. As for “membership”…it has an unfortunate connotation of exclusivity which is probably what we don’t want to impart:

1 April 1990 […]
The closed fplangc mailing list continued for committee discussions, but increasingly debate took place on the public Haskell mailing list. Members of the committee became increasingly uncomfortable with the “us-and-them” overtones of having both public and private mailing lists, and by April 1991 the fplangc list fell into disuse. All further discussion about Haskell took place in public, but decisions were still made by the committee.
[…]

1999-2002 In 1999 the Haskell Committee per se ceased to exist.

(A History of Haskell: Being Lazy With Class, page 5 of 55.)

Would “associate” be more appropriate for this purpose?