Great question!

• Haskell is pretty great!
• Companies seem to stop writing Haskell surprisingly often!
• Why does this happen?

An anecdote

About this talk

• Names don’t matter
• About Haskell-the-ecosystem rather than Haskell-the-language
• “All happy families are alike; each unhappy family is unhappy in its own way”
• In dialogue with Evan Czaplicki’s talk on Saturday
• My perspective on the absurdity of the tech industry during an especially absurd time

Suffering from success

Haskell is not mainstream

• Few domains where Haskell is the “safe” choice
• Nobody Ever Got Fired for Picking Java
• If your Haskell project fails, it’s Haskell’s fault

Haskell is not mainstream

Haskell is not trendy (anymore)

• Other, newer, languages: Rust, Lean, TypeScript, etc.
• Neither trendy enough nor niche enough

Haskell avoids “success at all costs”

• Very sensitive to any one entity exercising undue influence
• Hard to imagine an equivalent to e.g. Jane Street for OCaml
• Competing tools/philosophies, e.g. lens vs. optics

Companies seek to maximise profit

• Specifically shareholder value
• Often incompatible with highest quality product, personal development, and other desirable outcomes
• Worse Is Better

Most companies seek to maximise short-term profit

• Sometimes at the expense of long-term profit
• “The market can remain irrational longer than you can remain solvent” - Keynes
• Objectively bad decisions, e.g. doing a round of layoffs just before an earnings announcement to artificially boost the stock price

Companies undervalue maintenance

• Constant pressure to “do more with less”
• Nobody Ever Gets Credit for Fixing Problems that Never Happened
• Firefighting is more visible and more likely to be rewarded

Companies undervalue maintenance

Getting paid to write Haskell

1. Want to write Haskell
2. Convince someone with money to give you/your team that money
3. ????
4. PROFIT!!!

Some reasons you might use

• Believing that Haskell is a good fit for your problem
• Trying to hire skilled programmers
• Trying to retain a key employee
• Getting people to work on something boring

A non-exhaustive list

• The Second System
• The Experiment
• The Vanity Project
• The Giant Client
• The Acquisition
• The Insolvency
• The Ousting
• The Magpie
• The Rocket Explosion
• The Outgrowing

The situation

• Company decides to rewrite a non-Haskell codebase in Haskell
• Implied (or explicit) outcome is that the existing team will lose their jobs

The problem

• Adversarial relationship with current team
• Existing codebase provides business value so you have to work on it
• Second-system effect

Second-system effect

Hofstadter’s Law

It always takes longer than you expect, even when you take into account Hofstadter’s law.

Suggestions

• Ask yourself why a rewrite is happening
• Many successful Haskell adoptions look like sidecars (microservices?)
• Threatening someone’s livelihood is not a good way to get them on your side
• They probably don’t want to learn a new language
• Explore your options! (This will come up a lot)

The situation

• Company decides to rewrite a Haskell codebase in a non-Haskell language
• Implied (or explicit) outcome is that the existing Haskell team will lose their jobs
• Some might leave after the decision anyway

The situation

The problem

• All the problems I mentioned before
• Working Haskell product is on life support
• Maintaining it is not a fun job (because maintenance is undervalued!)

Suggestions

• Don’t panic!
• Hofstadter’s Law

The situation

• Funding (often fixed-term) to solve a problem with Haskell
• Eventually runs out

The problem

• Software projects and tight deadlines don’t mix
• Haskell is blamed for the failure
• The problem with venture capital is that you eventually run out of other people’s money

Suggestions

• Ride the gravy train till the end?
• Figure out an exit plan while things are still good

The situation

The situation

• Somebody starts a Haskell project as a way of enriching their personal brand

The problem

• Questionable attachment to Haskell or the project itself
• Increasingly bizarre requests for features

Suggestions

• Ride the gravy train until the end?
• Use the opportunity for personal/professional development
• Try not to leave a smoking crater behind

The situation

• Company ostensibly has multiple clients, in practice only one
• Client ends its contract

The problem

• Niche client pool
• Buyer’s market

Suggestions

• Try very hard to avoid this situation
• Exit plan!

The situation

The situation

• Company is acquired by another, larger, company
• Often an acqui-hire rather than a product acquisition

The problem

• Acquiring company has no attachment to Haskell
• One less Haskell company

Suggestions

• Stay just long enough for your stock to vest?
• Exit plan!

The situation

• Company runs out of money

The problem

• Company runs out of money

Suggestions

• You tell me?

The situation

• Influential person leaves the company or moves to a different role
• Their political capital was enabling the use of Haskell

The problem

• Replacement wants to “shake things up” which usually involves getting rid of Haskell
• see: Second System Pt. II

Suggestions

• This rarely happens immediately
• Could try to push back, but it might not work
• Exit plan!

The situation

The situation

• Company picked Haskell because it was trendy at the time but has pivoted to e.g. Rust

The problem

• Second System Pt. II
• No attachment to Haskell but also no attachment to the replacement language

Suggestions

• Ask the difficult questions about why technology choices are made
• Lindy effect

The rocket explosion

The situation

• Haskell codebase is too complicated

The problem

• Abstraction ceiling in Haskell is enormously high
• Second System effect means the replacement could be worse!

Suggestions

• Collectively agree and enforce which subset of Haskell you are using

The situation

• Company grows large enough that they have extremely specific requirements of the Haskell ecosystem

The problem

• Companies are reluctant to fund this development work or do it themselves
• “Avoid success at all costs” means the ecosystem isn’t going to do what you want just because you want it
• Eventually the codebase bitrots and a heroic effort to rewrite in another language starts

Suggestions

• Put your money where your mouth is (nobody ever gets credit for fixing problems that never happened)
• Most large non-Haskell companies have teams dedicated to developer tooling for other languages, why not us?

A non-exhaustive list (again)

• The Second System
• The Experiment
• The Vanity Project
• The Giant Client
• The Acquisition
• The Insolvency
• The Ousting
• The Magpie
• The Rocket Explosion
• The Outgrowing

Rarely Haskell-specific!

Take care of yourself first

• Your greatest leverage as an individual is your ability to walk away
• Your reputation outlives any job you will have

Just walk out

Spend your political capital wisely

• If the approval for Haskell flows from you, things will probably change when you leave
• Pointing out mistakes to management might not work

Collective action is pretty great

Collective action is pretty great

Failure is not the end

• Could do everything right and it might not matter
• That API DSL is very much alive and in use
• Make the most of the opportunities you have
• Hope springs eternal

Moving to New York City

Starting my SPJ selfie collection

Making new friends

Moving back to Australia

Scuba diving with my mum

SPJ selfie 2

Pursuing my passions

One more SPJ selfie

Slides

https://vaibhavsagar.com/presentations/how-to-stop-writing-haskell/