Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 .. 5022
— 502,152 events total
2020-11-22 23:48:16 <dminuoso> hekkaidekapus: I mean it's largely a management and manpower problem
2020-11-22 23:48:31 <merijn> UHC implements a handful, iirc
2020-11-22 23:48:52 → revprez_anzio joins (~revprez_a@pool-108-49-213-40.bstnma.fios.verizon.net)
2020-11-22 23:48:55 <dolio> A new report isn't going to happen by making average users type more crap at the top of their files.
2020-11-22 23:49:06 <dminuoso> hekkaidekapus: https://reasonablypolymorphic.com/blog/haskell202x/
2020-11-22 23:49:09 <hekkaidekapus> I’m not disputing the practical considerations of focusing on GHC right now.
2020-11-22 23:49:35 <dminuoso> hekkaidekapus: Different people, different team, different job.
2020-11-22 23:49:39 <dminuoso> It's not about "focusing on GHC"
2020-11-22 23:50:14 <merijn> Haskell' needs a cat herder :p
2020-11-22 23:51:02 <hekkaidekapus> If you are a compiler hacker in 2050, starting from a spec is far more interesting than reviving a 60 years old project.
2020-11-22 23:51:55 × arahael quits (~arahael@124-150-64-50.tpgi.com.au) (Quit: WeeChat 2.7.1)
2020-11-22 23:56:23 × livvy quits (~livvy@gateway/tor-sasl/livvy) (Ping timeout: 240 seconds)
2020-11-22 23:58:04 × conal_ quits (~conal@66.115.157.141) (Quit: Computer has gone to sleep.)
2020-11-23 00:00:10 → christo joins (~chris@81.96.113.213)
2020-11-23 00:00:15 × atk quits (~Arch-TK@ircpuzzles/staff/Arch-TK) (Quit: Well this is unexpected.)
2020-11-23 00:00:37 × m0rphism quits (~m0rphism@HSI-KBW-095-208-098-207.hsi5.kabel-badenwuerttemberg.de) (Ping timeout: 246 seconds)
2020-11-23 00:01:08 → conal joins (~conal@66.115.157.141)
2020-11-23 00:01:39 → atk joins (~Arch-TK@ircpuzzles/staff/Arch-TK)
2020-11-23 00:05:37 <hekkaidekapus> dminuoso: I read the page you linked. The main point there is a need of Benefactor Dictator For Life. Let’s say that I am a bit skeptic. ;)
2020-11-23 00:06:04 × raehik quits (~raehik@cpc95906-rdng25-2-0-cust156.15-3.cable.virginm.net) (Quit: WeeChat 2.8)
2020-11-23 00:06:44 <dminuoso> And funding.
2020-11-23 00:06:46 <dolio> I don't understand the argument. First off, it's inappropriate for a language standard to specify things down to an intermediate language level. That means every compiler has to make a ton of decisions exactly like GHC down to a very low level. If you just want documentation of what GHC does, there's already a bunch of papers and wikis and users guides that actually describe GHC already.
2020-11-23 00:07:36 → raehik joins (~raehik@cpc95906-rdng25-2-0-cust156.15-3.cable.virginm.net)
2020-11-23 00:08:16 × merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 240 seconds)
2020-11-23 00:08:35 <hekkaidekapus> dolio: If you check the backlog, you will see that I expressed a few hours ago that I am not fond of rigourous standards.
2020-11-23 00:09:21 <hekkaidekapus> My argument here is that taking over Haskell98 and Haskell2010 by a backdoor is a sneaky way of handling matters.
2020-11-23 00:12:33 × Tuplanolla quits (~Tuplanoll@91-159-68-239.elisa-laajakaista.fi) (Quit: Leaving.)
2020-11-23 00:12:48 <hekkaidekapus> As to why very detailed standard are not my cup of coffee, things end up set in stone, drastic changes practically impossible.
2020-11-23 00:13:41 → weilzach joins (~weilzach@pool-72-66-93-213.washdc.fios.verizon.net)
2020-11-23 00:13:48 <dminuoso> Unrelatedly, is there some cool trick to traverse over a data structure and fail if it has more than one target?
2020-11-23 00:13:59 × Lycurgus quits (~niemand@98.4.114.74) (Quit: Exeunt)
2020-11-23 00:14:09 <dminuoso> Beyond the very obvious toList and pattern matching on its length
2020-11-23 00:14:09 × weilzach quits (~weilzach@pool-72-66-93-213.washdc.fios.verizon.net) (Client Quit)
2020-11-23 00:14:11 <dolio> Also, if you want an active standard process, you need to make new viable Haskell implementations. The standard process is useful as a way of codifying the commonalities of multiple implementations. That's how it originated. The standards process being useless is a symptom of GHC being the only viable choice.
2020-11-23 00:15:29 → da39a3ee5e6b4b0d joins (~da39a3ee5@2403:6200:8876:6c06:c056:20b8:f8ee:6530)
2020-11-23 00:15:31 → weilzach joins (~weilzach@pool-72-66-93-213.washdc.fios.verizon.net)
2020-11-23 00:15:43 <dminuoso> dolio: Perhaps, it could arguably also help in that regard. If we had a new vision of what "standard haskell" in the current state of haskell could be, it would reduce the willingness to flip on extensions and stick to "standard haskell" - perhaps opening up doors for newer implementations
2020-11-23 00:15:50 <dminuoso> Right now hackage is extension wild west.
2020-11-23 00:16:30 <dolio> You mean, like specifying a bunch of standard GHC extensions that are considered good enough to be mainstream?
2020-11-23 00:16:35 <dolio> The 'sneaky' thing?
2020-11-23 00:16:40 <hekkaidekapus> heh
2020-11-23 00:16:57 <dminuoso> Not necessarily just that. But at least part of it, yes.
2020-11-23 00:17:07 × Fractalis quits (~Fractalis@2601:987:280:8d40:eda9:f9e1:2072:cea7) (Ping timeout: 260 seconds)
2020-11-23 00:17:27 × weilzach quits (~weilzach@pool-72-66-93-213.washdc.fios.verizon.net) (Read error: Connection reset by peer)
2020-11-23 00:17:27 <dminuoso> Haskell has resulted from academics doing research and trying out ideas. GHCs extensions are just that process.
2020-11-23 00:17:43 <dminuoso> I dont see anything fundamentally wrong as GHC extensions as precursors of haskell report features.
2020-11-23 00:18:33 → weilzach joins (~weilzach@pool-72-66-93-213.washdc.fios.verizon.net)
2020-11-23 00:18:37 <Axman6> unpopular opinion: extensions are just language feature flags #facebookdev
2020-11-23 00:19:10 <hpc> how else would you develop new language features, anyway?
2020-11-23 00:19:29 <Axman6> laborious committee process like C++
2020-11-23 00:19:43 × jb55 quits (~jb55@gateway/tor-sasl/jb55) (Ping timeout: 240 seconds)
2020-11-23 00:19:52 <hpc> even then you have an implementation of it somewhere before it gets standardized
2020-11-23 00:20:20 <hekkaidekapus> hpc: Let’s do that then.
2020-11-23 00:20:45 <hpc> and then make that implementation optional
2020-11-23 00:20:48 × vhs quits (~vhs@41.44.147.255) (Remote host closed the connection)
2020-11-23 00:20:56 <hpc> and then to make it easy to use, make it enabled with a comment at the top of the file
2020-11-23 00:20:59 <hpc> :P
2020-11-23 00:21:07 → vhs joins (~vhs@41.44.147.255)
2020-11-23 00:21:35 <hekkaidekapus> How many compilers are Report2010-compliants?
2020-11-23 00:22:28 <Axman6> probably up to one
2020-11-23 00:22:39 <hekkaidekapus> That’s a bummer.
2020-11-23 00:22:53 × alp quits (~alp@2a01:e0a:58b:4920:2d01:9668:71aa:c2be) (Ping timeout: 272 seconds)
2020-11-23 00:23:00 <hekkaidekapus> No standard process possible with a singleton.
2020-11-23 00:23:03 → phaul joins (~phaul@ruby/staff/phaul)
2020-11-23 00:25:56 <hekkaidekapus> stu002: I am not aware of such a Haddock feature.
2020-11-23 00:26:39 <hekkaidekapus> You can distribute the README by setting the `extra-files` in .cabal.
2020-11-23 00:26:45 × christo quits (~chris@81.96.113.213) (Remote host closed the connection)
2020-11-23 00:27:37 <hekkaidekapus> Some programmers make Haskell modules containing only documentation. Therein, the full extent of the Haddock syntax is available.
2020-11-23 00:28:00 <dolio> Anyhow, my point is, don't think about this in a backwards way. Forcing people to demand the release of another 'official standard' before they can type less garbage in their files isn't going to create more Haskell implementations. Because the implementations created the standard, not the other way around.
2020-11-23 00:29:16 <dminuoso> One prohibiting factor for new implementations to arise, is that without supporting most extensions, you couldn't use much of hackage. Transitively you quickly depend on a lot of extensions.
2020-11-23 00:29:48 <hekkaidekapus> dolio: Point taken.
2020-11-23 00:30:19 → arahael joins (~arahael@124-150-64-50.tpgi.com.au)
2020-11-23 00:30:38 <hekkaidekapus> stu002: For instance, this module has not a single Haskell identifier: <https://hackage.haskell.org/package/dhall-1.36.0/docs/Dhall-Tutorial.html>
2020-11-23 00:30:43 <dminuoso> And part of why so packages so freely depend on extensions without caring about interoperability, is because there's not even some principle Haskell2020 that you could even adhere to
2020-11-23 00:30:58 <dminuoso> The best we have is Haskell2010, and that's very limiting
2020-11-23 00:31:10 <dolio> Even if there was, there would be no reason to not use GHC's extensions beyond 2020.
2020-11-23 00:31:25 <dolio> Because there is no other compiler to use.
2020-11-23 00:31:32 → fendor_ joins (~fendor@91.141.2.69.wireless.dyn.drei.com)
2020-11-23 00:31:51 <dminuoso> I guess it's a vicious circle.
2020-11-23 00:32:49 → jb55 joins (~jb55@gateway/tor-sasl/jb55)
2020-11-23 00:33:29 × siwica quits (~user@p200300f6171ea70026a3ceac3db524b2.dip0.t-ipconnect.de) (Read error: Connection reset by peer)
2020-11-23 00:33:45 <monochrom> At this point, you can compare (another Haskell compiler competing with GHC) with (a new manufacturer of photocopiers competing with existing brands)
2020-11-23 00:33:52 × fendor__ quits (~fendor@77.119.131.102.wireless.dyn.drei.com) (Ping timeout: 246 seconds)
2020-11-23 00:33:58 <dminuoso> But I stand by my opinion that with a modern Haskell 202x standard, it would be easier to get people away from just using off-standard GHC extensions.
2020-11-23 00:34:11 <dminuoso> Without a standard you can't even try
2020-11-23 00:35:06 <monochrom> Users now have so much expectations of a Haskell compiler or a photocopier (now called "office document centre") that only GHC and only the existing very experienced manufacturers can live up to those expectations.
2020-11-23 00:35:14 → christo joins (~chris@81.96.113.213)
2020-11-23 00:35:51 → merijn joins (~merijn@83-160-49-249.ip.xs4all.nl)
2020-11-23 00:35:52 × jb55 quits (~jb55@gateway/tor-sasl/jb55) (Remote host closed the connection)
2020-11-23 00:36:07 <monochrom> Indeed look at the feature creeps into GHC that bring it further and further away from very simple Haskell, and the feature creeps that turned simple photocopiers to monstrous know-it-alls.
2020-11-23 00:36:13 → jb55 joins (~jb55@gateway/tor-sasl/jb55)
2020-11-23 00:36:41 <dolio> I don't think you need to compile hackage to be a useful Haskell compiler, incidentally. There's probably like 0.1% of hackage or less that matters.
2020-11-23 00:36:56 <dolio> Maybe less.
2020-11-23 00:37:34 <dminuoso> Ability to compile a notable portion of hackage is what a new implementation would require for good adoption.
2020-11-23 00:37:52 <dminuoso> Unless this was an industry implementation for some special use
2020-11-23 00:38:03 <dolio> PureScript compiles 0% of hackage, I expect.
2020-11-23 00:38:11 <dminuoso> PureScript is not Haskell though.
2020-11-23 00:38:22 <dolio> And yet people use it.
2020-11-23 00:38:43 <monochrom> But purescript compiles 100% of the purescript library repository :)
2020-11-23 00:39:11 <hekkaidekapus> dminuoso: That’s a tall order. Look at head.hackage and all the cycles going there to test new GHC features against a few Hackage packages.
←Prev  Next→
Page 1 .. 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 .. 5022

All times are in UTC.