Logs: liberachat/#haskell
| 2021-06-21 09:10:10 | → | yoctocell joins (~user@h87-96-130-155.cust.a3fiber.se) |
| 2021-06-21 09:10:43 | → | azeem joins (~azeem@176.200.214.195) |
| 2021-06-21 09:11:10 | <kuribas> | merijn: I also remember some article where they implemented search using a hybrid tree and linear search on arrays. |
| 2021-06-21 09:11:37 | <kuribas> | merijn: that was on CPU though. GPU's are another beast. |
| 2021-06-21 09:12:21 | × | MQ-17J quits (~MQ-17J@d14-69-206-129.try.wideopenwest.com) (Ping timeout: 265 seconds) |
| 2021-06-21 09:12:32 | → | hmmmas joins (~chenqisu1@183.217.200.246) |
| 2021-06-21 09:14:04 | × | hexo quits (~hexo@user/hexo) (Quit: Leaving) |
| 2021-06-21 09:14:07 | × | xsperry quits (~as@user/xsperry) (Ping timeout: 244 seconds) |
| 2021-06-21 09:15:13 | × | maroloccio1 quits (~marolocci@189.15.9.54) (Ping timeout: 246 seconds) |
| 2021-06-21 09:16:22 | <kuribas> | RE poor error handling in haskell, my PR to cassava for providing a decent error message (which should have been there), has been just sitting there for several weeks now. |
| 2021-06-21 09:17:14 | <kuribas> | This kind of sloppiness in core libraries just goes against all ideas of haskell being more rigorous. |
| 2021-06-21 09:17:56 | <kuribas> | It's like just an obsession for type level programming, GADTs etc, but then ignoring low hanging fruit. |
| 2021-06-21 09:18:10 | <ocramz> | cassava has been in a sorry state for a while now. I guess CSV is seen as low priority by the intelligentsia |
| 2021-06-21 09:18:43 | <ocramz> | (whereas the rest of the world lives inside excel tables and csv dumps) |
| 2021-06-21 09:18:49 | <kuribas> | ocramz: :-( We use CSV for everthing... |
| 2021-06-21 09:18:57 | → | MQ-17J joins (~MQ-17J@d14-69-206-129.try.wideopenwest.com) |
| 2021-06-21 09:18:59 | <ocramz> | I know, anybody doing data work uses csv |
| 2021-06-21 09:20:25 | <merijn> | If only there was some way to just take the existing work and improve it directly yourself... ;) |
| 2021-06-21 09:20:42 | <ocramz> | btw kuribas, I've had a good experience with csv-conduit ( https://hackage.haskell.org/package/csv-conduit ) for parsing , and serializing by hand with bytestring/text builders |
| 2021-06-21 09:20:53 | <merijn> | Then you could just pay programmers to make it to your liking, imagine that! |
| 2021-06-21 09:21:42 | <ocramz> | merijn, it's not about open source entitlement. we all do our share of OSS. but if a maintainer is not responsive and doesn't allow external help, then it's a problem |
| 2021-06-21 09:21:56 | <kuribas> | this |
| 2021-06-21 09:23:05 | <kuribas> | I am not even blaming the maintainer, he may be very busy or something, just saying it's a problem that reflects badly on the language. |
| 2021-06-21 09:23:24 | × | jneira quits (~jneira@212.8.115.226) (Quit: Client closed) |
| 2021-06-21 09:23:40 | → | jneira joins (~jneira@212.8.115.226) |
| 2021-06-21 09:23:46 | <merijn> | Right, but if the maintainer is too busy you ask if you can be co-maintainer and if not you fork? |
| 2021-06-21 09:24:43 | → | chronon joins (~chronon@user/chronon) |
| 2021-06-21 09:25:23 | <ocramz> | that would work if Hackage had a hierarchical name space, which is part of the social contract of our particular corner of open source |
| 2021-06-21 09:25:36 | <kuribas> | merijn: forking is possible, but not very convenient, and certainly doesn't reflect well on haskell. |
| 2021-06-21 09:25:56 | <ocramz> | e.g. kuribas.cassava is actively maintained, use that |
| 2021-06-21 09:25:59 | → | kenran joins (~kenran@b2b-37-24-119-190.unitymedia.biz) |
| 2021-06-21 09:26:14 | <merijn> | Sure it's not convenient, but what else do you propose? |
| 2021-06-21 09:27:09 | × | chexum quits (~chexum@gateway/tor-sasl/chexum) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | gnoo quits (~gnoo@user/gnoo) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | ChaiTRex quits (~ChaiTRex@user/chaitrex) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | bitdex quits (~bitdex@gateway/tor-sasl/bitdex) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | HotblackDesiato quits (~HotblackD@gateway/tor-sasl/hotblackdesiato) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | adanwan quits (~adanwan@gateway/tor-sasl/adanwan) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | gehmehgeh quits (~user@user/gehmehgeh) (Remote host closed the connection) |
| 2021-06-21 09:27:09 | × | hendursa1 quits (~weechat@user/hendursaga) (Remote host closed the connection) |
| 2021-06-21 09:27:13 | <kuribas> | I can ask to be co-maintainer... |
| 2021-06-21 09:28:27 | <merijn> | Like, these discussions are always about how lack of maintenance reflects on the library/community and how things "should" be professional, but in the end there's literally just 1 solution and that is "pay people enough to prioritise their open source work" |
| 2021-06-21 09:29:02 | <ocramz> | OR allow co-maintainers, as kuribas proposed one line above |
| 2021-06-21 09:29:21 | <merijn> | ocramz: I proposed that even earlier, but in effect that is still the same solution |
| 2021-06-21 09:29:33 | <merijn> | since, presumably, you are getting paid by your job to do that maintenance |
| 2021-06-21 09:29:57 | <ocramz> | many write OSS in their spare time |
| 2021-06-21 09:30:07 | <kuribas> | merijn: reviewing PRs shouldn't be that much work... |
| 2021-06-21 09:30:14 | → | __monty__ joins (~toonn@user/toonn) |
| 2021-06-21 09:30:19 | → | ChaiTRex joins (~ChaiTRex@user/chaitrex) |
| 2021-06-21 09:30:25 | <merijn> | ocramz: If you do, then you should realise why people might not have the time to prioritise |
| 2021-06-21 09:30:26 | <Taneb> | ocramz: if it's their spare time, they have no obligation on how to spend that time |
| 2021-06-21 09:30:31 | → | HotblackDesiato joins (~HotblackD@gateway/tor-sasl/hotblackdesiato) |
| 2021-06-21 09:30:33 | <merijn> | kuribas: Who cares if it's much work or not |
| 2021-06-21 09:30:35 | → | adanwan joins (~adanwan@gateway/tor-sasl/adanwan) |
| 2021-06-21 09:30:45 | × | hmmmas quits (~chenqisu1@183.217.200.246) (Quit: Leaving.) |
| 2021-06-21 09:31:00 | → | gehmehgeh joins (~user@user/gehmehgeh) |
| 2021-06-21 09:31:05 | <kuribas> | we should care for core libraries. |
| 2021-06-21 09:31:07 | <ocramz> | I feel we're talking past each other here |
| 2021-06-21 09:31:33 | <merijn> | Some of my stuff still doesn't support GHC 9.0, because after 1.5 year of stuck at home in lockdown I can't find the energy to work on it in my spare time |
| 2021-06-21 09:31:35 | → | eggplantade joins (~Eggplanta@2600:1700:bef1:5e10:945c:cf17:8af9:9d4a) |
| 2021-06-21 09:32:02 | <kuribas> | couldn't that be something for haskell consultancy companies? They get paid to sell haskell, so it is in their interest to have decent libraries. |
| 2021-06-21 09:32:19 | <ocramz> | since the Hackage namespace is write-only, and flat, IMO once you upload a new package it shoudl become part of the commons, and it should be easy by design to let new maintainers in, both in the repo and as hackage uploaders |
| 2021-06-21 09:32:35 | → | chexum joins (~chexum@gateway/tor-sasl/chexum) |
| 2021-06-21 09:32:44 | <ocramz> | without complex bargaining, or ego battles |
| 2021-06-21 09:32:46 | <merijn> | ocramz: That seems like a good way to get people to not upload code at all |
| 2021-06-21 09:32:49 | → | hendursa1 joins (~weechat@user/hendursaga) |
| 2021-06-21 09:32:55 | → | gnoo joins (~gnoo@user/gnoo) |
| 2021-06-21 09:33:20 | <merijn> | ocramz: I mean, you don't have to bargain, it's open source, you can simply fork it with a different name and use/advertise that? |
| 2021-06-21 09:33:25 | → | bitdex joins (~bitdex@gateway/tor-sasl/bitdex) |
| 2021-06-21 09:33:51 | <kuribas> | ocramz: wasn't their some drama a while ago about exactly that? Someone doing changes against the authors wishes. |
| 2021-06-21 09:34:01 | <ocramz> | because by forking we have two similar things, one of which is left to bitrot |
| 2021-06-21 09:34:19 | <ocramz> | kuribas : yeah |
| 2021-06-21 09:34:22 | <kuribas> | I thought there was already a process for abandoned projects. |
| 2021-06-21 09:34:22 | <Taneb> | kristjansson: cassava/Cassava? |
| 2021-06-21 09:34:48 | <Taneb> | kuribas* sorry |
| 2021-06-21 09:34:48 | <ocramz> | yeah |
| 2021-06-21 09:35:10 | <kuribas> | but I am not even talking about obscore/abandoned projects, I am talking about libraries considered core, like Cassava, Aeson, ... |
| 2021-06-21 09:35:21 | <ocramz> | this duplication is technically feasible but frankly sucks, because it shows that it was not possible to reach a civil agreement |
| 2021-06-21 09:35:21 | <merijn> | I mean, I wouldn't consider either of those core |
| 2021-06-21 09:35:36 | <kuribas> | merijn: what do you use for JSON then? |
| 2021-06-21 09:35:44 | <merijn> | kuribas: I don't use JSON |
| 2021-06-21 09:35:49 | × | eggplantade quits (~Eggplanta@2600:1700:bef1:5e10:945c:cf17:8af9:9d4a) (Ping timeout: 244 seconds) |
| 2021-06-21 09:35:53 | <kuribas> | merijn: good for you :) |
| 2021-06-21 09:35:59 | <ocramz> | kuribas: aeson, same thing |
| 2021-06-21 09:36:15 | <merijn> | Here's the key problem in "these core libraries", I doubt you'll find much agreement community wide on what is "core" |
| 2021-06-21 09:36:43 | <merijn> | Anything GHC depends on, sure. containers and vector also. I think people will already start disagreeing on unordered-containers |
| 2021-06-21 09:36:56 | <kuribas> | merijn: Cassava is the standard for CSV, Aeson for JSON, or am I wrong? |
| 2021-06-21 09:37:07 | <merijn> | kuribas: Define "standard" |
| 2021-06-21 09:37:10 | <ocramz> | merijn : a simple graph centrality argument willl show why cassava can be considered as "core" : https://packdeps.haskellers.com/reverse/cassava |
| 2021-06-21 09:37:18 | <kuribas> | merijn: as in, what everyone uses? |
| 2021-06-21 09:37:28 | × | elf_fortrez quits (~elf_fortr@adsl-64-237-239-58.prtc.net) (Quit: Client closed) |
| 2021-06-21 09:37:38 | <Taneb> | kuribas: that's community-accepted rather than standard. <blink> was supported by all commonly used browsers many years ago |
| 2021-06-21 09:37:53 | <merijn> | ocramz: I mean, that's 96 packages, most of which I never even heard off... |
| 2021-06-21 09:38:02 | → | pretty_dumm_guy joins (trottel@gateway/vpn/protonvpn/prettydummguy/x-88029655) |
| 2021-06-21 09:38:14 | <Taneb> | And, like, there are alternatives for both cassava and aeson, that do get used |
| 2021-06-21 09:38:15 | <ocramz> | "I haven't heard of a thing" is not a good argument why a thing is not important |
| 2021-06-21 09:38:37 | <merijn> | ocramz: I mean, vector has 1900 reverse dependencies |
| 2021-06-21 09:38:54 | <merijn> | 5300 for containers |
| 2021-06-21 09:39:06 | <merijn> | Those are orders of magnitude "more core" than cassava |
| 2021-06-21 09:39:16 | <ocramz> | all of these should be core, so what |
All times are in UTC.