Newest at the top
| 2024-05-14 12:26:56 +0200 | ubert | (~Thunderbi@2a02:8109:ab8a:5a00:d77a:cc4d:acdd:ee1e) |
| 2024-05-14 12:24:40 +0200 | _bo | (~bo@198.red-83-56-252.dynamicip.rima-tde.net) (Ping timeout: 260 seconds) |
| 2024-05-14 12:22:13 +0200 | bo_ | (~bo@198.red-83-56-252.dynamicip.rima-tde.net) |
| 2024-05-14 12:15:08 +0200 | <probie> | I wouldn't mind such behaviour to allow elision of names in a system where all arguments are named arguments |
| 2024-05-14 12:14:15 +0200 | titibandit | (~titibandi@user/titibandit) |
| 2024-05-14 12:06:56 +0200 | yin | (~yin@user/zero) |
| 2024-05-14 12:06:07 +0200 | <cheater> | you're right, hold up, let me implement a parser and a type checker real quick |
| 2024-05-14 12:05:30 +0200 | yin | (~yin@user/zero) (Ping timeout: 268 seconds) |
| 2024-05-14 12:00:21 +0200 | <ski> | but i don't want to discourage you from pondering it, if you want to |
| 2024-05-14 11:59:55 +0200 | <ski> | like, if you can elaborate this in a workable fashion, i'd be interested in taking a look .. i'm just saying i'm not seeing how it could work, in a way i'd consider reasonable |
| 2024-05-14 11:58:04 +0200 | random-jellyfish | (~developer@user/random-jellyfish) (Ping timeout: 260 seconds) |
| 2024-05-14 11:56:18 +0200 | lortabac | (~lortabac@2a01:e0a:541:b8f0:55ab:e185:7f81:54a4) (Quit: WeeChat 4.2.1) |
| 2024-05-14 11:55:09 +0200 | <ncf> | the unordered arguments thing sounds like it would also be garbage wrt type inference |
| 2024-05-14 11:53:16 +0200 | <ski> | yes |
| 2024-05-14 11:52:52 +0200 | <ski> | anyway, using records (or labelled multi-arity functions, if available) is one workable approach. what OCaml does, labelled curried arguments, also works |
| 2024-05-14 11:52:40 +0200 | <lortabac> | if you swap two arguments by mistake and it's obvious enough, an IDE may suggest inverting them |
| 2024-05-14 11:52:01 +0200 | <lortabac> | I think it would be nice to have this feature as an IDE action, but not baked in the language |
| 2024-05-14 11:51:52 +0200 | <danse-nr3> | one can also have f (default'{ arg = type }) |
| 2024-05-14 11:51:26 +0200 | <danse-nr3> | dunno, i just use product types and that gives me 99% of the functionality |
| 2024-05-14 11:51:21 +0200 | <ski> | danse-nr3 : <https://en.wikipedia.org/wiki/Hanlon's_razor> |
| 2024-05-14 11:51:10 +0200 | <lortabac> | you can select a record field by type if there is no ambiguity |
| 2024-05-14 11:50:54 +0200 | <lortabac> | IIRC generic-lens does something similar to what cheater is proposing |
| 2024-05-14 11:50:45 +0200 | <ski> | danse-nr3 : yea, that's basically that one |
| 2024-05-14 11:50:23 +0200 | <ski> | perhaps i'd understand it better, cheater, if you made a simple prototype type checker (and maybe parser, if you wish) |
| 2024-05-14 11:49:59 +0200 | <danse-nr3> | i remember occam's as "the simplest explanation is also the most likely" ... but it is lost in the mists of time. Maybe that was misreferenced |
| 2024-05-14 11:49:14 +0200 | <ski> | "Don't attribute to malice, what can be adequately explained by ignorance/stupidity." |
| 2024-05-14 11:48:51 +0200 | <danse-nr3> | v |
| 2024-05-14 11:48:46 +0200 | <ski> | no, it's not the same thing |
| 2024-05-14 11:48:39 +0200 | <danse-nr3> | huh i knew it as occam's razor... or is it different? |
| 2024-05-14 11:48:24 +0200 | <ski> | i genuienly don't know how you are envisioning this to work |
| 2024-05-14 11:48:02 +0200 | <cheater> | no |
| 2024-05-14 11:47:56 +0200 | <ski> | Hanlon's razor |
| 2024-05-14 11:47:39 +0200 | <cheater> | stop being confused on purpose |
| 2024-05-14 11:47:23 +0200 | <ski> | `x',`y',`z',`a',`b',`c' here are meta-variables, standing for arbitrary expressions. there's no a-priori associatiation of these, to the types `X',`Y',`Z',`A',`B',`C' .. if you mean just actual variable names `x',`y',`z',`a',`b',`c' .. then that's very limited, doesn't scale, isn't really composable |
| 2024-05-14 11:46:27 +0200 | <cheater> | it was all part of the same example |
| 2024-05-14 11:46:10 +0200 | <cheater> | it was an example in the context of what i said before about the type signature |
| 2024-05-14 11:45:53 +0200 | <ski> | not in the application, you did not |
| 2024-05-14 11:45:42 +0200 | <cheater> | i specifically typed out their types ski |
| 2024-05-14 11:45:35 +0200 | <danse-nr3> | still with this stuff? What is wrong with my suggestion to use a product type? |
| 2024-05-14 11:45:23 +0200 | <ski> | what if the types of `x',`y',`z' are the same as the types of `a',`b',`c' (respectively) ? |
| 2024-05-14 11:45:23 +0200 | <cheater> | at that point it's easy to figure out which piece goes in which hole |
| 2024-05-14 11:45:04 +0200 | <cheater> | i'm talking about the fact that most often with functions in haskell you either have arguments that all have different types, or just a small amount amount of arguments of equal types, which can be represented as a tuple (or, uncurried, with -> ) |
| 2024-05-14 11:44:48 +0200 | <cheater> | i'll disagree with that last one |
| 2024-05-14 11:44:25 +0200 | <ski> | (at least not, if you want the language to be sane) |
| 2024-05-14 11:44:05 +0200 | <ski> | (can't reaally provide them, out of order, without labelling them) |
| 2024-05-14 11:43:49 +0200 | <cheater> | yes. not what i'm talking about |
| 2024-05-14 11:43:41 +0200 | <ski> | it's labelled arguments, which you can provide out of order |
| 2024-05-14 11:43:16 +0200 | <cheater> | yeah that's not really what i'm talking about |
| 2024-05-14 11:40:46 +0200 | <ski> | cheater : <https://ocaml.org/manual/5.2/lablexamples.html> |
| 2024-05-14 11:38:23 +0200 | deriamis | (deriamis@ec2-54-187-167-69.us-west-2.compute.amazonaws.com) |