Newest at the top
| 2024-12-28 03:55:21 +0100 | <homo> | tcl offers a clean syntax to quickly prototype gui with tk |
| 2024-12-28 03:53:46 +0100 | <homo> | it's when such abstractions are separated in different languages they are easy to read and write |
| 2024-12-28 03:53:08 +0100 | <homo> | go's abstraction is CSP-style concurrency, you don't need library for that, make's abstraction is dependency management, again you don't need separate library for that, prolog's abstraction is search, once again no library is needed for that |
| 2024-12-28 03:53:00 +0100 | <fp> | I'm trying to write a parser with megaparsec and I want to read in input up to a delimiter. The delimiter can be a few different characters, but I'd also like for it to be EOF. However, the char parsing is of type =Parser Char=, but EOF is =Parser ()=. How can I make a single parser that capture both characters and EOF |
| 2024-12-28 03:50:25 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 248 seconds) |
| 2024-12-28 03:49:30 +0100 | <homo> | many languages are wonderful when they offer thinking from different perspective not available in other languages |
| 2024-12-28 03:48:33 +0100 | <Sisyphean> | but I could be wrong. |
| 2024-12-28 03:48:23 +0100 | <Sisyphean> | homo, makes sense. I think that'd be ideal. I still think there's always going to be naturally many languages and we just have to deal with it. |
| 2024-12-28 03:46:27 +0100 | <Sisyphean> | yeah I stand corrected mid-80s |
| 2024-12-28 03:46:24 +0100 | <homo> | plan9 is haskell of operating systems |
| 2024-12-28 03:46:04 +0100 | <homo> | unix was born in the 70s, plan9 is not that old |
| 2024-12-28 03:45:56 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2024-12-28 03:45:44 +0100 | <homo> | Sisyphean anyway, I didn't mean to offend you, I have personal struggle with multi-paradigm languages because of how unpleasant their syntax looks to the point it's necessary to use auto-completion plugins in the editor and how hard it is to understand someone else's code, instead it's much easier to read when there are separate languages for every paradigm, I recommend to watch "public static void" and "simplicity is complicated" talks by Rob Pike |
| 2024-12-28 03:44:56 +0100 | <Sisyphean> | or am I crazy? |
| 2024-12-28 03:44:44 +0100 | <Sisyphean> | what does that have to do with today? |
| 2024-12-28 03:44:38 +0100 | <Sisyphean> | wasn't that in the 70s? |
| 2024-12-28 03:44:30 +0100 | <Sisyphean> | plan 9 |
| 2024-12-28 03:34:50 +0100 | Sisyphean | (~Sisyphean@user/Sisyphean) Sisyphean |
| 2024-12-28 03:33:10 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
| 2024-12-28 03:32:53 +0100 | Sisyphean | (~Sisyphean@user/Sisyphean) (Read error: Connection reset by peer) |
| 2024-12-28 03:32:16 +0100 | Zenen | (~Zenen@node-1w7jr9uk45pdylasbk2itgwn9.ipv6.telus.net) (Quit: Client closed) |
| 2024-12-28 03:32:09 +0100 | <homo> | s/where/whether/ |
| 2024-12-28 03:31:43 +0100 | <homo> | hurd might be the only exception, as they say hurd's translators are same thing as plan9's file servers, but I don't know anything about hurd besides it's being microkernel, and I don't know where hurd offers network-transparent namespaces |
| 2024-12-28 03:30:50 +0100 | <geekosaur> | no, I'm talking about a hacked-up port to *BSD back in the 1990s |
| 2024-12-28 03:29:11 +0100 | <homo> | it works only for tools internal to plan9port |
| 2024-12-28 03:28:58 +0100 | <homo> | do you mean a hack-up used in plan9port? |
| 2024-12-28 03:28:46 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2024-12-28 03:28:40 +0100 | <geekosaur> | opening windows, drawing graphics, etc. |
| 2024-12-28 03:28:30 +0100 | <homo> | with 9P, namespaces and files as REST API resources there is too much flexibility to do anything without ffi |
| 2024-12-28 03:28:29 +0100 | <geekosaur> | I recall everything under it being done by echoing escape sequences |
| 2024-12-28 03:28:29 +0100 | emmanuelux | (~emmanuelu@user/emmanuelux) (Quit: au revoir) |
| 2024-12-28 03:28:15 +0100 | <geekosaur> | someone did a hack-up at one point |
| 2024-12-28 03:26:57 +0100 | <homo> | plan9's window system cannot be ported to unix because unix doesn't have network-transparent namespaces |
| 2024-12-28 03:26:25 +0100 | <geekosaur> | you're reminding me of porting plan9's window system to sco unix 🙂 |
| 2024-12-28 03:25:41 +0100 | <homo> | do you want haskell to render stuff on the screen? open file /dev/draw and write draw commands |
| 2024-12-28 03:25:00 +0100 | <homo> | I actually want to play with haskell on plan9, because there ffi doesn't make sense, all you have to do is open a file and start reading and writing to that file |
| 2024-12-28 03:23:49 +0100 | <homo> | I was about to rewrite my message and accidentally hit enter :) |
| 2024-12-28 03:23:03 +0100 | <geekosaur> | for a minor example, in Haskell's FFI all values must be forced and all constructors removed because C can't deal with them, and conversely constructors must be added to C values |
| 2024-12-28 03:22:46 +0100 | <homo> | and consistency is about working with other projects |
| 2024-12-28 03:21:58 +0100 | <geekosaur> | the problem with subset of features is you may find yourself limited to the libraries that support the same interface, unless the compiler and runtime are very clever about being able to combine the features at the RTS level |
| 2024-12-28 03:21:45 +0100 | <haskellbridge> | <loonycyborg> I still think it's mostly matter of taste |
| 2024-12-28 03:21:11 +0100 | <homo> | Sisyphean that "subset" is the reason I prefer using different languages so that code looks very clean to the eyes, every time I open up .java, .rs, .cpp in the editor I think I am getting eye cancer because of how ugly they look |
| 2024-12-28 03:20:55 +0100 | <geekosaur> | PL/I was IBM's attempt at an "everything" programming language, originally unifying FORTRAN and COBOL and later adding SNOBOL and other languages. It was a huge, bloated nightmare |
| 2024-12-28 03:20:18 +0100 | <haskellbridge> | <loonycyborg> Like singletons and libs based on them very much remind me boost.mpl and other c++ template metaprogramming libs |
| 2024-12-28 03:19:53 +0100 | <Sisyphean> | or are you implying some kind of inconsistency (which I don't get since the language must compile) |
| 2024-12-28 03:19:35 +0100 | <Sisyphean> | or are you saying I'm incapable of that and most users need to be spoonfed? |
| 2024-12-28 03:19:20 +0100 | <Sisyphean> | homo: why can't I choose the subset of features that solves my problem? |
| 2024-12-28 03:19:18 +0100 | <haskellbridge> | <loonycyborg> Caused by features that are arguably missing |
| 2024-12-28 03:19:08 +0100 | <haskellbridge> | <loonycyborg> and both have some large amount of emergent paradigms. |
| 2024-12-28 03:18:10 +0100 | <haskellbridge> | <loonycyborg> * variable at runtime, |