Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Haskell Elephant in the Room

stephendiehl.com

Post reply on HN
Print view View original

71–80 of 341 posts

Re: The Haskell Elephant in the Room

#71

Post by CyberDildonics »

Barrin92 wrote: post #46
ogogmad wrote: post #39

Earlier quoted context omitted.

How do you know what's "pseudo" and not "pseudo" in economics? Given that it isn't a science. I'm genuinely curious.

Economics is a science. You get an idea what's 'pseudo' or not by working in or familiarizing yourself with the field. Say, The New neoclassical synthesis is very much as standard as it gets, whereas Marxist or Austrian economics exist on the fringes. In general something is pseudo-scientific if it operates outside of the formalisms or tools of that particular discipline, especially if it pretends that it does not.

Economics is not something that can be tested easily on a large scale, so most of it becomes about trying to explain why things happened in retrospect.

Re: The Haskell Elephant in the Room

#72

Post by pron »

I see this problem with formal methods, which are also increasingly used in the cryptocurrency world. I compare their use there to the claim that a tightrope is safe because it's anchored to towers assembled from cards made of the strongest titanium alloy.

Having said that, I doubt the impact of Haskell's reputation on cryptocurrency users, technical or otherwise. The myth that Haskell results in more correct programs might still be alive in portions of the Haskell community despite the failure to support the claim with any evidence, but few outside that community have ever heard of that myth, let alone believe it.

Re: The Haskell Elephant in the Room

#73

Post by dgellow »

LargeWu wrote: post #38
markhollis wrote: post #25

Earlier quoted context omitted.

I'm not an expert and I don't do value judgments, but I think he's talking about something called 'smart contracts'. Haskell would make it easier to verify smart contracts. I thought he was talking about use cases like those as Cardano and with "some of the very founding contributors" he meant Philip Wadler (amongst others). EDIT: didn't want to imply that Cardano is a shady company.

No, he's not talking about the technical features of blockchain. He's talking about the general shadiness of a nontrivial number of actors in the cryptocurrency sector itself, which have had a propensity for deceptive marketing, and even outright fraud. His concern is that if Haskell gets the reputation of being beholden to these interests it will make the Haskell ecosystem undesirable to legitimate actors.

But... who actually thinks this way when they select technologies? If Go was used for a lot of crypto scam, I wouldn’t spend a second thinking about it when I’m deciding to use it for my non-scam, non-crypto’s project.

I have difficulty to understand how it is an issue that a programming language is used for a niche that has bad reputation.

Re: The Haskell Elephant in the Room

#74

Post by 1tCKV3QfIo »

twat wrote: post #8
hannofcart wrote: post #3

What is it about Haskell that makes it a hot candidate for use in cryptocurrency applications? Can someone here shed some light? Read the article but it mostly seems to be a diatribe against crypto on philosophical grounds. I wish it were meatier on the technical reasons for this alleged relationship between Haskell and crypto industry.

Haskell always struck me as a language you'd want to use to feel superior to the people who code in JS and python, and crypto always struck me as a field you'd want to work in to feel superior to the people who work in industries that actually generate profits. Also, another possible reason is that many crypto people tend to confuse complexity with ingenuity. While the bitcoin whitepaper tries to make a complex topic…

Just say you don't want to learn more than the absolute minimum when it comes to programming language theory. No one will give you shit if you are honest about it.

Re: The Haskell Elephant in the Room

#76

Post by analyte123 »

bondarchuk wrote: post #11

Very much a generic anti-cryptocurrency rant, and very little haskell-specific clarifications. I understand not wanting to name specific projects (Cardano comes to mind; " Cardano is a blockchain platform built on the groundbreaking Ouroboros proof-of-stake consensus protocol, and developed using the Haskell programming language: a functional programming language that enables Cardano to pursue evidence-based developm…

This is an article about a giant pot of questionably earned money buying influence in a software ecosystem that he's been a big part of. But this is hardly the first time for this, in other cases the pot of questionably earned money came from selling people's data for ads or from Windows licenses.

The question in all of these cases is what specific bad influences the money can have on the software ecosystem and how community standards and governance can mitigate them. I don't really see a lot of answers in this article besides ill-defined ethical compromise of developers. I skimmed the book he mentioned (The Politics of Bitcoin: Software as Right-Wing Extremism) which sounds interesting but it doesn't really make its own case very well: half of the "extremist" citations are from anonymous online commenters and you can't go 2 paragraphs without straight up name-calling and ham-fisted guilt by association.

Re: The Haskell Elephant in the Room

#77

Post by rmrfrmrf »

At this point, the entire financial sector is decoupled from actual productivity, and to pass moral judgement of one highly exploitative industry over another solely due to one's ability to be regulated seems myopic, at best.

Re: The Haskell Elephant in the Room

#78

Post by CyberDildonics »

bodhiandpysics1 wrote: post #57
hannofcart wrote: post #3

What is it about Haskell that makes it a hot candidate for use in cryptocurrency applications? Can someone here shed some light? Read the article but it mostly seems to be a diatribe against crypto on philosophical grounds. I wish it were meatier on the technical reasons for this alleged relationship between Haskell and crypto industry.

There is a certain class of computer programs that can best be described as a really big table. Basically, an input taken from a finite set, you get an output taken from a finite set. This type of program is really easy to write in Haskel because you can prove that both the conplete input set and output set are covered using the type system. In other words you can prove that you deal with every possible case (though…

That's basically a DAG - directed acyclical graph. Data is transformed from inputs to outputs with no state or branching on the macro level of the transform. Lots of programs have parts that map to this very well. The cracks show when someone realizes this and thinks it is a silver bullet to build an architecture that ONLY does this. Then the parts that inevitably do need branching, state, and complex loops become a big problem. Combined with resource management (which can be thought of as mixing in branching and state) and the simplistic approach that seemed like a silver bullet turns into a nightmare once the realities of real software set in.

Re: The Haskell Elephant in the Room

#79

Post by tempodox »

prionassembly wrote: post #34

> "what economists call non-productive assets" That's not a technical term nor a "term of art" in economics. Google Scholar returns results from management and accounting journals, as well as some pseudo-economics ("heterodox") pamphlets.

https://www.nasdaq.com/articles/what-are-productive-assets-2...
Post reply on HN