03:19 <discocaml> <dubious245> Utop was the first time I really used a repl to experiment with a language.

03:20 <discocaml> <dubious245> I just used the python repl as a calculator cause the kde calc app was borked last time I tries to use it. :mbDerp:

07:40 <hanker> what's the debugging support for Lwt code?

10:05 <discocaml> <mbs> '

11:16 <discocaml> <empathy.monster> Same here. I'll admit that it did take me a little while to get used to, but tbh now it's hard for me to mess with new language without one, and I used to couldn't stand repls. I've yet to find one as good as utop though.

13:49 <discocaml> <youngkhalid> ghci for me 🥲

14:23 <discocaml> <dubious245> I almost tried haskell first but something distracted me from giving it a go, I had everything installed I just didn't do anything with it.

14:24 <humasect> i came from haskell to ocaml, looking for "middle point" between C and super-high level

14:24 <humasect> learned a lot. might go back

14:44 <discocaml> <youngkhalid> God saved you from suffering. Not sure you would've liked its strict purity.

14:47 <discocaml> <youngkhalid> But as @humasect0156 says, you learn a lot from it. It's good for learning the functional paradigm (that's how I knew about it). That language is like a good old teacher, very harsh but it's worth it.

14:47 <discocaml> <dubious245> I am just so demotivated working in c++ I'd rather do ANYTHING else.

14:48 <discocaml> <youngkhalid> C++ has its flaws as well, I'd use it for medium-sized projects, but the language has sooooo much stuff in it that it can quickly become a hell maintaining/debugging..etc.

14:49 <discocaml> <youngkhalid> Unless you're very careful with all the power that comes with it / work with highly skilled people in a large codebase.

14:58 <humasect> youngkhalid: saved me? talking to me? about haskell? ...

14:58 <discocaml> <youngkhalid> I answered to @ Dubious message

14:58 <humasect> works nice. but, it just shifts the power from the programmer to the compiler.

14:59 <humasect> ok

14:59 <humasect> either type more or think more

14:59 <discocaml> <youngkhalid> sorry ?

15:01 <discocaml> <darrenldl> i think discord "reply to" dont carry over to irc

15:01 <discocaml> <youngkhalid> Yes , I supposed this

15:01 <discocaml> <youngkhalid> But what does "thinking more" has to do with it ?

15:02 <discocaml> <darrenldl> ah, hm

15:03 <discocaml> <darrenldl> nvm, i thought it was confusion due to messages not appearing as directed replies over at irc

15:06 <discocaml> <dubious245> "either type more or think more" What?

15:07 <hanker> am i the only one who really likes ocaml for the "compiling to js" deal

15:07 <hanker> like you have this native compilation language and you can also use it in browsers

15:07 <hanker> blows my mind

15:08 <humasect> couldn't be simpler. type more, or think more

15:08 <humasect> C type type type Haskell think think think

15:09 <discocaml> <dubious245> Ooooh, I get it now.

15:09 <hanker> dubious245 yeah I thought type as in type systems too

15:10 <hanker> i find ocaml is the sweet point between haskell and C too

15:13 <humasect> ah oops sry =)

15:13 <humasect> yeah

15:13 <humasect> and the "syntax tricks" as i like to call them, feels like im being smart =) C does same

15:14 <humasect> but_this_kind_of_syntax iPreferThisKind

15:17 <hanker> i kinda prefer snake case

15:21 <discocaml> <youngkhalid> @humasect Ah I get it, you were talking about Haskell. I thought you was talking to me in an imperative way xD

15:24 <discocaml> <youngkhalid> And yeah, it's exactly that. In Haskell you have to think more than you type. That's the cost of having syntactically cleaner and more pure code

15:55 <humasect> youngkhalid: ah yeah =) and yeah. some call it "pencil+papering"

15:55 <humasect> still need a scratchpad

16:41 <hanker> i think another thing i like a lot with ocaml

16:42 <hanker> the language goes way out of its way to avoid making you name variables you don't actually want to name because you'd give them generic useless names

16:42 <hanker> like you can decompose records pretty much anywhere. you have function and other syntactic sugar

16:56 <discocaml> <contificate> hanker: no, I really like compiling OCaml to JS as well - I honestly do not believe TypeScript etc. really address the problems of JavaScript in any meaningful sense, so it's nice to be able to write business logic in a capable language (along with being able to use normal tooling like menhir, ocamllex, etc.)

16:57 <discocaml> <contificate> the problem with Haskell is the illusion of "patterns arising naturally" disappears once you start gluing things together - people start dealing with idioms like "reader over IO" and monad transformer stacks

16:58 <discocaml> <contificate> and, if you're like me, you'll question what you get in return for bending over backwards so often - and the answer is nothing but partly-theoretical suggestions from Haskellers about referential transparency for reasoning - impossible when you obfuscate everything in tedious ways anyway

17:31 <dh`> there's nothing natural about the ways normal program logic appears in monad constructions

17:35 <dh`> e.g. it's kind of cute that you can do "when (x < 3) $ do [sub-block]" and when can just be a function and doesn't need to be a builtin

17:35 <dh`> but it's of no practical value, and because builtins have syntax they tend to be easier to read

17:36 <dh`> (furthermore, the errors you get out of the compiler if you leave off the () or the $ are quite exotic)

18:09 <discocaml> <five.sided.fistagon> monad transformers were where my haskell journey ended, indeed