| --- Log opened Mon Jul 22 00:00:39 2019 |
| 00:43 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:10 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 01:14 | | Vash [Vash@Nightstar-sjaki9.res.rr.com] has quit [[NS] Quit: Leaving] |
| 05:05 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 05:05 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:12 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:16 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:29 | | Kindamoody is now known as Kindamoody|afk |
| 13:44 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has joined #code |
| 13:44 | | mode/#code [+o celticminstrel] by ChanServ |
| 13:50 | | celticminstrel is now known as celmin|away |
| 14:53 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 14:53 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:59 | | ErikMesoy1 [Bruker@Nightstar-idn3a6.bb.online.no] has joined #code |
| 16:01 | | ErikMesoy [Bruker@Nightstar-idn3a6.bb.online.no] has quit [Ping timeout: 121 seconds] |
| 16:25 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 17:00 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 17:12 | | ErikMesoy1 is now known as ErikMesoy |
| 19:08 | <&jeroud> | Today has not been a very good day. |
| 19:09 | <&Derakon> | Oh no. |
| 19:11 | <&jeroud> | I know this because I am at band practice trying to think about music instead of being angry about how so many people seem okay with the garbage fire that is so much of the software the modern world is built on. |
| 19:11 | <&Derakon> | If it's any consolation, the bad software we write is mostly just a reflection of the bad software we think with. |
| 19:12 | <&Derakon> | So it's not particularly surprising that people are OK with it. |
| 19:19 | <~Vornicus> | I guess that is an advantage of doing math software |
| 19:19 | <&jeroud> | In my experience (I have written a lot of terrible software in my time) the cause is mostly a combination of "we didn't actually think about the implications of doing this" and "we'll just crap out a minimal-effort thing and move on". |
| 19:19 | <~Vornicus> | I can technically prove the correctness of my algorithm |
| 19:20 | <~Vornicus> | ...too bad it is as yet incorrect~ |
| 19:21 | <&jeroud> | I know that I am in the absolute worst part of the industry for this stuff. |
| 19:23 | <&jeroud> | But then I look at golang and npm and a dozen other things millions of software developers use every single day and I weep. |
| 19:31 | <&McMartin> | Worse is better, etc |
| 19:33 | <&jeroud> | No, worse is worse. |
| 19:34 | <&jeroud> | It's only "better" if you're buried in the toxic morass of "Silicon Valley" tech culture. |
| 19:34 | <&McMartin> | Uh |
| 19:34 | <&McMartin> | I am not sure if you have actually missed my referent |
| 19:34 | <&jeroud> | I don't think I have. |
| 19:34 | <&McMartin> | Because if you hadn't you might have instead connected it to New Jersey. |
| 19:35 | <&McMartin> | "A specific talk by Richard Gabriel and ultimately Lisp's excuse for why C ate its lunch" |
| 19:35 | <&McMartin> | Which he goes back on several times later |
| 19:36 | <&McMartin> | I have in the past had reasons for thinking he is wrong but I have not evaluated them recently. |
| 19:37 | <&jeroud> | I wasn't entirely aware of the whole origin story, but I do know that in the past decade or so that idea has been embraced and extended by a significant chunk of the industry. |
| 19:38 | <&McMartin> | If I were talking to basically anyone else here I would suggest that they were incorrectly guessing the content of the talk |
| 19:39 | <&McMartin> | Your favored language however is Elixir and this is not entirely necessarily true |
| 19:40 | <&McMartin> | But I think it should be uncontroversial to note that npm is bad in ways that C is not, and that many of the ways C is bad are often unironically features. |
| 19:40 | <~Vornicus> | Is he complaining that C Moves Fast And Breaks Things? |
| 19:41 | <~Vornicus> | (I don't know the talk) |
| 19:41 | <&McMartin> | No |
| 19:42 | <&McMartin> | He's complaining that it is theoretically possible that Unix syscalls can return EINTR. |
| 19:42 | <&McMartin> | http://dreamsongs.com/WIB.html |
| 19:44 | <&McMartin> | Certain aspects of this talk are definitely now outdated, and outdated thoroughly enough that the things that made it outdated are *themselves* outdated |
| 19:44 | <&McMartin> | However, I have not investigated them thoroughly enough to see if ASDF and Quicklisp end up being as bad as npm in practice~ |
| 19:46 | <&McMartin> | Common Lisp had itself not yet finished standardization at the point of this article |
| 19:47 | <&McMartin> | Also Scheme must have been R4RS at best at this point |
| 19:48 | <~Vornicus> | I remember being... |
| 19:48 | <~Vornicus> | ... befuddled at the number of things common lisp had builtins for? |
| 19:48 | <~Vornicus> | like it felt like hundreds of things and they were just there |
| 19:49 | <&jeroud> | The thing is, the problems lisp had were relatively big difficult things. |
| 19:49 | <&McMartin> | Yaeh |
| 19:49 | <~Vornicus> | Which seemed weird to me, coming from -- at the time -- a background of python and C where if I wanted a thing I'd load it specifically because having it all already there meant your program had too much shit going on |
| 19:50 | <&McMartin> | The argument for "Worse is Better" is actually "If you think 'worse is better' your metrics are probably wrong" |
| 19:50 | <~Vornicus> | and then of course the first thing i tried to use it for was a thing that lisp was totally unsuited to do in the first place and I should've been using PostScript~ |
| 19:50 | <&McMartin> | C in 1989 kicked the absolute everliving *shit* out of every existent Lisp in 1989. |
| 19:52 | <&McMartin> | The CL implementations I have found are technically capable of producing performant code, but getting it to run on anybody else's computer is an exercise in frustration and pointless siloing, even though everyone supports EFFI these days |
| 19:52 | <&jeroud> | The problems npm and friends have are lots of little thing that would have been obvious and fairly straightforward to solve or avoid if the people building them had taken the time to think things through. |
| 19:53 | <&jeroud> | Some of this is just ignorance and lack of concern for other people's experiences. |
| 19:53 | <&McMartin> | Right, while the "system calls might be interrupted by hardware that needs to kick you back to user mode" problem was one that was never satisfactorily solved |
| 19:54 | <&McMartin> | "Even the case of Unix vs. Multics misses an important point. While Multics may have been a better operating system, Unix had the twin advantages of actually existing, and running on a wide variety of fairly cheap hardware." |
| 19:55 | <&jeroud> | Things like "if the network is slow and/or flakey, npm will take half an hour to continue doing a bunch of now-useless work and will then fail in a manner that does not allow easy recovery". |
| 19:55 | <&jeroud> | This is a problem I had with npm multiple times. |
| 19:58 | <&jeroud> | There seemed to be no desire on the part of the maintainers to do anything to remedy this behaviour. |
| 20:02 | | Degi [Degi@Nightstar-as9iq6.dyn.telefonica.de] has joined #code |
| 20:04 | <&McMartin> | Yeah, that is an objection that is not really in the realm of the original worse-is-better argument |
| 20:05 | <&McMartin> | And if I'm in a bad mood, once you start with "I'd like to use JavaScript as a full-scale application or server development language" you have already committed to all kinds of unnecessary pain |
| 20:06 | <&McMartin> | So the closest point of contact to the argument is "once you narrow the universe of discourse to the people who have already committed to doing this what is the best available technology" |
| 20:06 | <&McMartin> | But part of worse-is-better's alleged superior survival characteristics is that it can actually be improved after it becomes popular due to its actual existence |
| 20:07 | <&McMartin> | And it seems like they've been falling down on that~ |
| 20:10 | <&jeroud> | Yup. |
| 20:11 | <&McMartin> | But thanks to its open source nature, anybody can totally just go improve it and thus suborn its popularity! |
| 20:11 | <&McMartin> | ^_^ |
| 20:11 | <&McMartin> | ^_^ |
| 20:11 | <&jeroud> | I guess what I'm complaining about is the huge segment of the industry that heard the phrase "worse is better" and then used it as an excuse to stop caring about quality. |
| 20:12 | <&jeroud> | Speaking of which (actually not speaking od which at all), I now need to spend most of an hour driving home in the rain. |
| 20:12 | <&McMartin> | Hooray |
| 20:13 | <&McMartin> | Anyway, yeah, thinking about it more |
| 20:14 | <&McMartin> | Someone from the Erlang tradition would consider the approach that paper calls "The Right Thing" to be impossibly deluded and possibly precision-engineered to make its practitioners irrelevant |
| 20:15 | <&McMartin> | Since the stereotype there is "certain kinds of things Literally Never Go Wrong" |
| 20:15 | <&McMartin> | Erlang's approach broadly speaking seems to be "the fact that you are alive meant certain kinds of things didn't go wrong, so you don't have to do the work of explicitly testing for that. Someone else can worry about whether or not you are alive." |
| 20:23 | <&Derakon> | The lack of quality in software can readily be seen to be driven by two simple factors: 1. software development is hard, in unpredictable ways. 2. software development is expensive. |
| 20:24 | <&Derakon> | Why would you devote an unknown but guaranteed to be large effort towards unquantifiable "craftsmanship"? |
| 20:24 | <&Derakon> | There's so many other more productive uses of your time and money, even if you're working for free. |
| 20:25 | <&Derakon> | You need pretty damn strong counterarguments for why quality is important, if you're going to overcome the natural impulse to say "this is good enough" pretty much as soon as the software meets the base need. |
| 21:01 | < ErikMesoy> | I can't help thinking "base need" is a flexible term doing a lot of work in that statement. |
| 21:39 | <&McMartin> | Yes, and the differing definition of "base need" is IMO the primary distinction between the MIT and New Jersey styles |
| 21:51 | <&McMartin> | Today in "People crazier than me": http://cowlark.com/2019-07-20-cowgol-agc/ |
| 21:53 | <&Derakon> | "asynchronous logic mostly implemented with discrete logic gates wired together by hand" aaaaaaah |
| 21:57 | <&McMartin> | Yeah, uh |
| 21:57 | <&McMartin> | ICs literally had not been invented yet |
| 21:57 | <~Vornicus> | "hand weave into core rope memory" |
| 21:58 | <&McMartin> | It was this project that drove their development |
| 21:58 | <&McMartin> | six-pin TTL ICs are cutting-edge 1970s tech sirs |
| 21:59 | <&Derakon> | Yeah, this is mostly an "I know people did this but I don't want to do this" kind of thing. |
| 21:59 | <&Derakon> | Some parts of our past should stay in our past. :p |
| 21:59 | <&McMartin> | Also core memory was third-generation memory technology~ |
| 22:20 | | gnolam [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 22:20 | | mode/#code [+o gnolam] by ChanServ |
| 22:50 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 23:11 | | Derakon [Derakon@Nightstar-fr5qel.ca.comcast.net] has quit [Connection closed] |
| 23:46 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| --- Log closed Tue Jul 23 00:00:41 2019 |