Hacker News Reader

Comment by eckza | Hacker News Reader

> no industrial success

Because FP is hard, and because it's easier to get 1000 Python monkeys to shit out half-baked garbage than it is to get 10 Haskell developers to agree on a set of blessed language extensions.

But when you get it right, your shit stops chewing a hole in itself in prod, and you can sleep at night and leave Slack alerts off over the weekend.

It's fine. Y'all can have your half-baked tire fires. "Industrial success" is a broken metric for "whether or not something is going to create real value for your product". We'll just keep shipping. And sleeping.

View on HN

The idea of our craft is to deliver value, not to find a new weird way to do the same thing. If dentists were programmers, the most expensive ones would be removing teeth via the rectum.

> The idea of our craft is to deliver value, not to find a new weird way to do the same thing. If dentists were programmers, the most expensive ones would be removing teeth via the rectum.

What a nice quote this will make :-)

PM me the name of who I should attribute this too.

What happens when your service scales and you get a performance bottleneck? Can you.. connect to a prod instance and look at the number of allocated instances? Can you do VM-event streams with barely any overhead continuously in prod? Will you actually be able to circumvent the performance bug?

Don’t get me wrong, I like Haskell and it is a more than fine language. Nonetheless, it doesn’t have anywhere close to the tooling Java has, and lazyness/FP is quite unintuitive regarding performance. And pure FP in itself is not guaranteed to result in better results by any form of evidence we have today, and doesn’t defend against logic bugs in any shape or form over what Java can do.

Also, it’s very unprofessional to look down on software engineers using different languages then you do..

> Also, it’s very unprofessional to look down on software engineers using different languages then you do..

it's a major red flag for me when giving or taking interviews.

Agreed. Most of the hard problems in software engineering are completely separate from language and stack choice. Hyper-focus on these things smells of expert-beginner.

> But when you get it right, your shit stops chewing a hole in itself in prod, and you can sleep at night and leave Slack alerts off over the weekend.

Okay, when FP people get it right, they can let the rest of us know and we'll jump on the bandwagon.

Until they get it right[1], let the enterprise devs do what they need to without pouring scorn on them.

[1] For whatever meaning of "right" you have.

I just want to stop for a moment and admire some of your metaphor pileups:

> monkeys to shit out half-baked garbage

> your shit stops chewing a hole in itself

> half-baked tire fires

Awesome. Terrifying, but awesome.

> We'll just keep shipping. And sleeping.

When do you actually ship though?

I’ve seen plenty of half-baked garbage from scala FP purists. Bad code is bad code.

Reply on news.ycombinator.com