| HN Mirror

Y	Hacker News new \| ask \| show \| jobs

by benreesman 710 days ago

Haskell is a really nice language. In general I don’t identify as an X programmer for any value of X: I tend to write in a half dozen languages daily and they all suck in their own special way.

But on two separate occasions I made important career decisions with opportunity cost to work with highly lethal GHC contributors: those people are just really good.

If Haskell sucks like all languages it’s because Haskell excels at using computers to compute something: Haskell considers data shuffling a strictly secondary concern compared to doing actual computations.

2 comments

random3 710 days ago

How do you distinguish data shuffling from computation? What’s actual computation from this perspective?

link

mrkeen 710 days ago

Before I was good at Haskell, I would approach a data-processing job sequentially based on the next thing that needs to be done.

I want to open a file, and I can't read it all at once, so I'll use a FileReader and it should be buffered, so I'll wrap it with a BufferedReader. I'll use try-with-resources and click into the classes because I can't remember if the contract of the outermost reader is that it will close the inner readers too.

Right, now I'll grab the next n bytes from the stream, and start thinking about the algorithm. Swear a bit when I think about crossing the buffer boundaries, and on-and-on...

The IO concerns are very much interwoven with the algorithm.

In Haskell I just start by writing one function from bytes to bytes. That's the computation. Then when that's done I expose that function as bytes to bytes.

Others can hook it up to files, webservers, pipe it through gzip, whatever!

link

lucianbr 710 days ago

Reading a row from a database and putting it on the screen, and reading some numbers from the keyboard and putting them in the database. These things I would not call computation. I mean sure, displaying needs to compute coordinates for where to light up pixels, but that's all already written. I just call it. Same with updating btrees when writing to the db.

I'm guessing if all you do is this kind of db - screen - keyboard and back stuff, haskell is not very useful, if not actively a hindrance.

link

benreesman 709 days ago

Haskell is actively a hindrance if one is mostly moving bytes from one place to another: the only thing that matters when you need to talk to 7 databases each different is fashion. The language that has bindings to all 7 each with a zillion users is the one you should use.

If you’re moving a lot of undifferentiated bytes the language you should use is historically C, more recently C++ (which is still the standard), or maybe soon Rust (which looks to become the standard).

If IO is a small part of your problem, performance needs to be good but not insane, and you’re mostly thinking about algorithms and mathematics?

Haskell is a very pragmatic choice there. OCaml is strong here too, and TypeScript is a very cool compromise between “mainstream” and “we do math here”.

link

tossandthrow 710 days ago

Philosophically speaking there is no difference.

What parent commenter probably refers to is that you think in terms of computations and not in terms of data units.

And that is just tremendously elegant.

link

082349872349872 710 days ago

Philosophically speaking there's a great difference.

Data shuffling doesn't —in principle— lose information; computation does. ("evaluation is forgetting")

In https://news.ycombinator.com/item?id=32498382 "glue code" and "parsley code" are data shuffling, while "crunch code" is computation.

link

tossandthrow 709 days ago