Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 462 463 464 465 466 467 468 469 470 471 472 .. 5022
502,152 events total
2020-10-06 12:57:37 oxide joins (~lambda@unaffiliated/mclaren)
2020-10-06 12:57:47 <dminuoso> quazimod1: It might take 6 months before you can write solid code though, if you're completely new to Haskell.
2020-10-06 12:57:55 <sm[m]> Haskell is the language you can be sure you can refactor yourself way out of wrong decisions
2020-10-06 12:58:05 <dminuoso> The curve is difference, you can't just "simply switch". Overall Id say its worth learning
2020-10-06 12:58:10 <quazimod1> rails was good to me, to be honest. in fact ruby is the easiest language i've ever used and they're adding static typing to it so it'll be waaay more scalable soon. Shame it's so slow
2020-10-06 12:58:12 <sm[m]> s/way//
2020-10-06 12:58:24 <maerwald> sm[m]: also depends
2020-10-06 12:58:29 <kuribas> yeah, I never understood the argument why clojure would be more "organic", as if it automatically grows with your problems...
2020-10-06 12:58:40 <dminuoso> quazimod1: The type system is absolute rubbish. :)
2020-10-06 12:58:47 <quazimod1> dminuoso: it's a type system at least
2020-10-06 12:58:50 <dminuoso> It's not really
2020-10-06 12:58:59 <quazimod1> i don't need it to be brilliant, i need it to stop devs making type mistakes
2020-10-06 12:59:03 <maerwald> sm[m]: vertical refactoring isn't much easier (as in: change the way you express effects, exceptions etc)
2020-10-06 12:59:14 <dminuoso> The main problem is, Ruby by design and ecosystem promotes reflection techniques and monkey patching so hard
2020-10-06 12:59:16 <maerwald> horizontal yes
2020-10-06 12:59:18 <quazimod1> dminuoso: have you inspected the rails 3 type system?
2020-10-06 12:59:18 <dminuoso> The type system cant keep track of it
2020-10-06 12:59:27 × albert_99 quits (~Albert@p200300e5ff0b5b39340e476afbaec452.dip0.t-ipconnect.de) (Ping timeout: 260 seconds)
2020-10-06 12:59:32 <dminuoso> quazimod1: Ive stopped with Ruby 2 years ago in favor of Haskell
2020-10-06 12:59:34 <quazimod1> yeah the monkey patching isn't great
2020-10-06 12:59:42 <sm[m]> quazimod1: yesod is a good industrial strength "rails" worth looking at - not as easy to use but faster
2020-10-06 12:59:44 <kuribas> quazimod1: ruby (and python/perl/clojure/php...) are easy tot get started with, but hard to maintain.
2020-10-06 13:00:02 <quazimod1> sm[m]: in 10 years i've never had speed issues with rails tbh
2020-10-06 13:00:04 <sm[m]> you might also like to chat out the new IHP
2020-10-06 13:00:04 <dminuoso> There's very few popular languages that I consider to have respectable type systems. Rust is one of them.
2020-10-06 13:00:16 <kuribas> thanks to monkey patching, you never know what your standard library functions are doing.
2020-10-06 13:00:18 mirrorbird joins (~psutcliff@2a00:801:2d5:9d73:ff00:6553:d451:a276)
2020-10-06 13:00:21 <sm[m]> quazimod1: ok, I thought you just said it was slow
2020-10-06 13:00:24 × ryansmccoy quits (~ryansmcco@193.37.254.27) (Ping timeout: 258 seconds)
2020-10-06 13:00:33 <quazimod1> sm[m]: for computation & the like, web server is fine
2020-10-06 13:00:46 ryansmccoy joins (~ryansmcco@193.37.254.27)
2020-10-06 13:01:12 <quazimod1> dminuoso: i'm 90% settled on writing this thing in rust actually
2020-10-06 13:01:19 <quazimod1> that language does not appear to fuck around
2020-10-06 13:01:27 <quazimod1> the tooling is easy to use too, as far as i've used it
2020-10-06 13:01:45 <maerwald> kuribas: what does ("/foo" </> "/bar") (from System.FilePath) return?
2020-10-06 13:02:02 <quazimod1> but it was in reading about stuff that i stumbled on the lazy v strict thing... such a shame ocaml is surrounded by very ... non commercial thinking contributors
2020-10-06 13:02:05 <kuribas> maerwald: "/foo/bar"?
2020-10-06 13:02:07 <maerwald> no
2020-10-06 13:02:23 urodna joins (~urodna@unaffiliated/urodna)
2020-10-06 13:02:31 <kuribas> "/foo//bar" then? :-)
2020-10-06 13:02:39 <quazimod1> sm[m]: i don't think i would do any financial computation stuff in ruby, even ruby 3 with the JIT
2020-10-06 13:02:44 <maerwald> kuribas: "/bar"
2020-10-06 13:02:47 rihards joins (~rihards@balticom-142-78-50.balticom.lv)
2020-10-06 13:02:51 <kuribas> ??
2020-10-06 13:02:54 <maerwald> :)
2020-10-06 13:02:59 albert_99 joins (~Albert@p200300e5ff0b5b39340e476afbaec452.dip0.t-ipconnect.de)
2020-10-06 13:03:25 <kuribas> FilePath is a historical smell
2020-10-06 13:03:26 Tops2 joins (~Tobias@dyndsl-095-033-093-242.ewe-ip-backbone.de)
2020-10-06 13:03:26 <dminuoso> quazimod1: The one thing that Haskell is just pretty good at, reasoning about code like equational reasoning.
2020-10-06 13:03:27 <sm[m]> quazimod1: I build a financial app in Haskell, its great
2020-10-06 13:03:36 <maerwald> kuribas: https://github.com/haskell/filepath/issues/49
2020-10-06 13:03:42 <quazimod1> sm[m]: tell me more
2020-10-06 13:03:44 <dminuoso> For financial code, Haskell would be at the top of my list probably :)
2020-10-06 13:03:51 <quazimod1> dminuoso: it appears to me that that's the case too
2020-10-06 13:03:53 <maerwald> kuribas: just saying... are you *sure* we are doing better wrt sdlib :p
2020-10-06 13:04:14 <dminuoso> The better you can reason about code, the more likely your code is correct, the easier you can maintain your code, and it makes refactoring that much easier.
2020-10-06 13:04:24 <quazimod1> dminuoso: yeah i guess i got spooked by the lazy thing. It doesn't help that the jane street nerds specifically didn't use haskell because it would make their high speed trading less predictable... which spooked me again
2020-10-06 13:04:28 <quazimod1> even though i'm nothing near high speed
2020-10-06 13:04:30 <sm[m]> smooth evolution from "I have no idea what I'm doing" to "pretty decent architecture that gets the job done"
2020-10-06 13:04:34 <kuribas> maerwald: at least it's consistently wrong :)
2020-10-06 13:04:39 <sm[m]> over 13 years..
2020-10-06 13:04:45 <Uniaika> sm[m]: :D
2020-10-06 13:04:48 <dminuoso> quazimod1: One of Haskell's primary use is actually in financial institutions.
2020-10-06 13:05:10 <dminuoso> It's not as if Haskell was "made for it" or that its "particularly well suited", but it *is* used there.
2020-10-06 13:05:19 <sm[m]> quazimod1: (hledger.org)
2020-10-06 13:05:32 <quazimod1> i can understand why, if i was going to analyse existing data it'd be different
2020-10-06 13:05:44 <quazimod1> i'm sure lazy or not the throughput would be fantastic & the code very very reliable
2020-10-06 13:05:50 <dminuoso> It felt like every third person I met in London as Haskell eXchange was working for some financial institution
2020-10-06 13:05:51 <sm[m]> but yes look at cards I for a much more significant financial app
2020-10-06 13:06:13 <quazimod1> also whoever mentioned clojure before: i do not need a language that encourages deveolpers to dump their brain thoughts into files thank you
2020-10-06 13:06:18 <sm[m]> Cardano. Dangit, phone..
2020-10-06 13:06:31 drbean joins (~drbean@TC210-63-209-203.static.apol.com.tw)
2020-10-06 13:06:33 <quazimod1> i will never implement another thing in lisp again
2020-10-06 13:06:36 <quazimod1> except emacs
2020-10-06 13:06:53 <quazimod1> sm[m]: hey that's cool
2020-10-06 13:07:08 × notzmv quits (~user@unaffiliated/zmv) (Ping timeout: 260 seconds)
2020-10-06 13:07:29 <dminuoso> quazimod1: But yeah, dont be spooked by lazyness. Overally I think it makes writing code easier, and the rare space leaks it brings with it can be debugged - for some reason some people trip into space leaks more often than others.
2020-10-06 13:07:33 <frdg> now that I think about it, Haskell isn't good for talking about sets because sets should be able to hold different types of elements.
2020-10-06 13:07:38 cosimone joins (~cosimone@2001:b07:ae5:db26:b248:7aff:feea:34b6)
2020-10-06 13:07:47 <dminuoso> frdg: Why should it?
2020-10-06 13:08:03 <quazimod1> dminuoso: so what do i do, use rust, python, ocaml or haskell for the quant app
2020-10-06 13:08:28 berberman_ joins (~berberman@123.118.97.97)
2020-10-06 13:08:32 <dminuoso> quazimod1: Why are you asking #haskell what you should do?
2020-10-06 13:08:38 <dminuoso> That seems like the wrong question
2020-10-06 13:08:42 <quazimod1> you're all pretty smart
2020-10-06 13:08:50 <frdg> dminuoso: [1,2,foo] is valid mathematically but in haskell this is a type error.
2020-10-06 13:08:51 <sm[m]> prototype in all until you see a winner!
2020-10-06 13:08:52 <ski> Scheme is cool. CL seems a bit less principled, more focused on getting things done, now. haven't looked at Clojure
2020-10-06 13:08:53 <dminuoso> Im not very smart
2020-10-06 13:08:55 × berberman quits (~berberman@2408:8207:2563:5ae0:584e:a9ff:fe9b:d3fe) (Ping timeout: 244 seconds)
2020-10-06 13:08:58 <dminuoso> Which is why Im using Haskell
2020-10-06 13:08:59 <quazimod1> never had any jackassery here that i can remember
2020-10-06 13:09:14 <dminuoso> It allows me to write good code *despite* being only average.
2020-10-06 13:09:28 <sm[m]> heh same :)
2020-10-06 13:09:32 <quazimod1> ski: CL is an abomination
2020-10-06 13:09:35 <dminuoso> Of course people here will tell you how good Haskell is
2020-10-06 13:09:41 <dminuoso> So Im not quite sure what you hope to hear.
2020-10-06 13:09:43 <ski> frdg : "[1,2,foo] is valid mathematically" -- in actual mathematical practice, that doesn't happen (or only very rarely)
2020-10-06 13:09:43 <quazimod1> lol
2020-10-06 13:09:52 <quazimod1> i'm honestly thinking rust

All times are in UTC.