Logs: freenode/#haskell
| 2020-09-17 16:18:58 | <carter> | i get add and mult confused |
| 2020-09-17 16:19:04 | <carter> | sum, pair, & and par |
| 2020-09-17 16:19:23 | <carter> | i dont really see the metaphor for the additive vs mult binary operators perspective |
| 2020-09-17 16:19:24 | <ski> | hm. when i read Wadler's paper on session types and linear logic, i was thinking all connectives were the opposite of what i expected |
| 2020-09-17 16:19:47 | <carter> | and i alwasy forget which things are additives vs mult |
| 2020-09-17 16:19:57 | <ski> | (iow, replaced by duals, everywhere. or possibly it was just for one of the subsystems described) |
| 2020-09-17 16:20:20 | <carter> | i dont like a lott of / most expositions i've seen |
| 2020-09-17 16:20:45 | <dolio> | ski: Might depend on how they were being used, I guess. |
| 2020-09-17 16:20:45 | <ski> | consider `a * (b + c) = (a * b) + (a * c)' |
| 2020-09-17 16:20:53 | <ski> | `*' is multiplicative, `+' is additive |
| 2020-09-17 16:20:55 | <dolio> | Inputs vs outputs. |
| 2020-09-17 16:20:57 | <ski> | that's how i remember it |
| 2020-09-17 16:21:45 | <ski> | dolio : yes, presumably, but it still seemed simpler to me, to just flip all the connectives, then. not upset people's expectations, for what seemed to be no reason |
| 2020-09-17 16:23:10 | <ski> | carter : fwiw, i don't like the usual notation for (some of) the connectives .. |
| 2020-09-17 16:23:16 | → | bicho_rastrero joins (~cerdito@154.85-87-39.dynamic.clientes.euskaltel.es) |
| 2020-09-17 16:23:24 | <dolio> | Which one? :) |
| 2020-09-17 16:23:24 | <ski> | .. so, in personal notes, i'm using my own convention :) |
| 2020-09-17 16:23:31 | <ski> | yea, there's several |
| 2020-09-17 16:23:39 | <ski> | i had Girard's one in mind |
| 2020-09-17 16:23:58 | <dolio> | I think both are wrong categorically. |
| 2020-09-17 16:24:27 | <ski> | (i reconstructed it from memory, after getting exposed to the ideas. then, after a bit, i noticed my reconstructed version was not the same as in the papers i read) |
| 2020-09-17 16:24:42 | <carter> | dolio: which are wrong how? |
| 2020-09-17 16:25:04 | <dolio> | carter: Additives are (co)product, but I think neither notation uses * for additve conjunction. |
| 2020-09-17 16:26:08 | <dolio> | That would suggest + for additive disjunction, too, but maybe Girard doesn't do that, either. |
| 2020-09-17 16:26:17 | <dolio> | It's a circled +? |
| 2020-09-17 16:26:41 | <dolio> | I've heard Girard's notation is supposed to remind you of distributive laws. |
| 2020-09-17 16:26:48 | → | nan` joins (~nan`@unaffiliated/nan/x-5405850) |
| 2020-09-17 16:27:18 | <carter> | lol |
| 2020-09-17 16:27:46 | <dolio> | But I'm not sure that's worth it. |
| 2020-09-17 16:29:26 | × | p-core quits (~Thunderbi@koleje-wifi-0046.koleje.cuni.cz) (Quit: p-core) |
| 2020-09-17 16:29:28 | × | jedws quits (~jedws@101.184.189.58) (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 2020-09-17 16:29:45 | <ezzieyguywuf> | so pretty much always just use ghcup rather thn whatever my OS's package manager has, right? |
| 2020-09-17 16:29:45 | → | p-core joins (~Thunderbi@koleje-wifi-0046.koleje.cuni.cz) |
| 2020-09-17 16:29:57 | → | Gurkenglas joins (~Gurkengla@unaffiliated/gurkenglas) |
| 2020-09-17 16:30:10 | → | carlomagno joins (~cararell@inet-hqmc02-o.oracle.com) |
| 2020-09-17 16:30:34 | <maerwald> | ezzieyguywuf: if your OS PM gets it right... then that's also nice |
| 2020-09-17 16:30:47 | <ski> | clearly ⌜+⌝ ought to be coproduct ("either","internal choice"), ⌜×⌝ be product ("with","external choice"), ⌜0⌝ be initial object ("void"), ⌜1⌝ be terminal object ("erase"), i thought. that leaves ⌜⊕⌝ for multiplicative disjunction ("par"), ⌜⊗⌝ for multiplicative conjunction, and then i took ⌜⊥⌝ ("bot") and ⌜⊤⌝ ("top","unit") as the respective neutral elements for those |
| 2020-09-17 16:30:47 | <dumptruckman> | So... If I wanted to model a game state a Monad would be sensible for this, right? |
| 2020-09-17 16:31:00 | <maerwald> | ezzieyguywuf: but ghcup is faster with adding new versions, most likely |
| 2020-09-17 16:31:17 | <dolio> | Oh yeah, the units are all mixed up, too. |
| 2020-09-17 16:31:28 | <ezzieyguywuf> | maerwald: they did the ghc nicely, they just pulled the binary from upstream, but the cabal install is a mess - I end up having to jump through all these hoops to get the cabal dependencies installed in the PM |
| 2020-09-17 16:31:56 | <ezzieyguywuf> | I was planning to just install ghc and cabal from the OS, then cabal install to grab things I need for dev and install the locally |
| 2020-09-17 16:32:57 | hackage | uniqueness-periods-vector-filters 0.1.1.0 - A library allows to change the structure of the 'RealFrac' function output. https://hackage.haskell.org/package/uniqueness-periods-vector-filters-0.1.1.0 (OleksandrZhabenko) |
| 2020-09-17 16:33:17 | <ski> | yes, Girard uses : ⌜⊕⌝,⌜&⌝,⌜0⌝,⌜⊤⌝ for the additives, and ⌜⅋⌝,⌜⊗⌝,⌜⊥⌝,⌜1⌝ for the multiplicatives |
| 2020-09-17 16:34:01 | × | p-core quits (~Thunderbi@koleje-wifi-0046.koleje.cuni.cz) (Remote host closed the connection) |
| 2020-09-17 16:34:25 | × | coot quits (~coot@37.30.52.181.nat.umts.dynamic.t-mobile.pl) (Quit: coot) |
| 2020-09-17 16:34:27 | <ski> | dumptruckman : monads can be nice for keeping track of state, yes |
| 2020-09-17 16:35:40 | <ski> | (oh, forgot to say that mult. conj. is "(both) .. and .."/"tensor") |
| 2020-09-17 16:38:16 | → | coot joins (~coot@37.30.52.181.nat.umts.dynamic.t-mobile.pl) |
| 2020-09-17 16:38:20 | × | fendor quits (~fendor@t204-126.demo.tuwien.ac.at) (Remote host closed the connection) |
| 2020-09-17 16:41:30 | <Uniaika> | phadej: do you think if SqlQQ can be safely extracted from postgresql-simple? I'd like to use it for sqlite too but I'm not familiar with postgresql-simple's internals |
| 2020-09-17 16:43:08 | <ezzieyguywuf> | k so really...no definitive answer on OS vs ghcup |
| 2020-09-17 16:43:09 | <ski> | dumptruckman : some alternative might perhaps be to try to keep local state, maybe via some kind of stream processors or something (FRP ?) |
| 2020-09-17 16:43:14 | → | LKoen joins (~LKoen@81.255.219.130) |
| 2020-09-17 16:43:17 | <dumptruckman> | mmk |
| 2020-09-17 16:43:37 | <ezzieyguywuf> | any gentoo users in here? |
| 2020-09-17 16:43:44 | <dumptruckman> | How would I lift the list out of this to be able to use intersect on an instance of Shares? `data Shares = Shares [Sharable]` |
| 2020-09-17 16:43:46 | <maerwald> | ezzieyguywuf: why |
| 2020-09-17 16:44:00 | <maerwald> | also: gentoo is not the distro that does get GHC right |
| 2020-09-17 16:44:12 | <ski> | dumptruckman : pattern-matching |
| 2020-09-17 16:44:19 | <dumptruckman> | what would that look like |
| 2020-09-17 16:44:28 | hackage | pusher-http-haskell 2.0.0.1 - Haskell client library for the Pusher Channels HTTP API https://hackage.haskell.org/package/pusher-http-haskell-2.0.0.1 (willsewell) |
| 2020-09-17 16:44:35 | <dumptruckman> | do i need to define `intersect :: Shares -> Shares -> Shares`? |
| 2020-09-17 16:44:38 | <ski> | intersectShares :: Shares -> Shares -> Shares |
| 2020-09-17 16:44:39 | <dumptruckman> | ok |
| 2020-09-17 16:44:46 | <monochrom> | "OS vs ghcup" depends on the OS. However, for 90% of OSes, go with ghcup. |
| 2020-09-17 16:44:52 | <dumptruckman> | oh, i have to name it differently? |
| 2020-09-17 16:45:03 | <ski> | instersectShares (Shares ss0) (Shares ss1) = Shares (intersect ss0 ss1) |
| 2020-09-17 16:45:15 | <ski> | depends on whether you import `intersect' qualified |
| 2020-09-17 16:45:37 | <dumptruckman> | hmm ok |
| 2020-09-17 16:45:59 | → | geekosaur joins (42d52102@66.213.33.2) |
| 2020-09-17 16:46:47 | <dumptruckman> | is it silly to use the & operator for this? |
| 2020-09-17 16:46:53 | <dumptruckman> | is that even possible? |
| 2020-09-17 16:46:58 | <monochrom> | You can. |
| 2020-09-17 16:47:00 | <ezzieyguywuf> | maerwald: there's been changes in gentoo rather recently, the haskell situation is lots better |
| 2020-09-17 16:47:20 | <ezzieyguywuf> | but I still think I'm going to stick to ghcup - I don't want to have, like, mtl and stuff installed at the system level |
| 2020-09-17 16:47:30 | × | coot quits (~coot@37.30.52.181.nat.umts.dynamic.t-mobile.pl) (Quit: coot) |
| 2020-09-17 16:47:50 | <maerwald> | ezzieyguywuf: I'd be surprised, because the gentoo PM cannot handle installing the same package mulitple times |
| 2020-09-17 16:48:01 | <maerwald> | which you need for haskell packages |
| 2020-09-17 16:48:51 | <maerwald> | The proper way to packages haskell is: only package GHC and cabal and *nothing* else (no haskell packages) |
| 2020-09-17 16:49:17 | <ski> | (dumptruckman : btw, in Haskell terminology, you wouldn't talk about "an instance of Shares", but rather about values of (type) `Shares') |
| 2020-09-17 16:49:20 | <maerwald> | everything else ends up showing limitations of the package manager with the only exception being nix |
| 2020-09-17 16:49:32 | <maerwald> | but then... you know... you'd have to use nix, ugh |
| 2020-09-17 16:50:18 | <ezzieyguywuf> | maerwald: it can, it's called "slots" |
| 2020-09-17 16:50:49 | <ezzieyguywuf> | maerwald: that's what I wanted to do, but in the current setup, I have separate packages for cabal's dependencies, i.e. mtl, fail, etc. |
| 2020-09-17 16:51:04 | <maerwald> | ezzieyguywuf: no, I'm talking about the same version |
| 2020-09-17 16:51:09 | <maerwald> | slots don't do that |
| 2020-09-17 16:51:18 | <ezzieyguywuf> | also, on a different note - I accidentally ran `ghcup` in `/root` as `myUser`, and it gave me a permission error when it tried to do `cabal new-update` |
| 2020-09-17 16:51:21 | <phadej> | Uniaika: https://hackage.haskell.org/package/postgresql-simple-0.6.2/docs/src/Database.PostgreSQL.Simple.SqlQQ.html#sql just copy it |
| 2020-09-17 16:51:27 | <phadej> | it doesn't really do anything smart |
| 2020-09-17 16:51:29 | <dumptruckman> | ski: thanks, i wasn't sure about that |
| 2020-09-17 16:51:31 | <ezzieyguywuf> | shouldn't it always dereference my ~ variable, no matter what the current directory is? |
| 2020-09-17 16:51:53 | <phadej> | (cliking source on hackage docs pages is the trick, often enough ;) |
| 2020-09-17 16:51:54 | <ezzieyguywuf> | maerwald: so there's a need for the same version of the same package to be installed multiple times? |
| 2020-09-17 16:52:06 | <maerwald> | ezzieyguywuf: You cannot install filepath-1.2.0.0 for multiple GHC slots on gentoo |
| 2020-09-17 16:52:21 | <maerwald> | that's a fact, slots don't solve this |
| 2020-09-17 16:52:22 | <maerwald> | nix does |
| 2020-09-17 16:52:38 | <ski> | dumptruckman : it's fine, got what you meant. just pointing it out, for more efficient future communication |
| 2020-09-17 16:53:02 | <ezzieyguywuf> | neat! is ghcup tui using the boxes thing? |
| 2020-09-17 16:53:08 | <ezzieyguywuf> | urm, bricks |
All times are in UTC.