| --- Log opened Mon Apr 03 00:00:38 2017 |
| 00:07 | | Turaiel[Offline] [Brandon@Nightstar-8t8.2pu.55.45.IP] has quit [Ping timeout: 121 seconds] |
| 00:08 | | Turaiel[Offline] [Brandon@Nightstar-8t8.2pu.55.45.IP] has joined #code |
| 00:08 | | mode/#code [+o Turaiel[Offline]] by ChanServ |
| 00:08 | <&McMartin> | Right, for my immediate purposes the Guido Note Server is *exactly* what I want. |
| 00:50 | | Derakon [Derakon@Nightstar-5mvs4e.ca.comcast.net] has quit [Connection closed] |
| 00:51 | | Derakon [Derakon@Nightstar-5mvs4e.ca.comcast.net] has joined #code |
| 00:51 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 01:57 | | Jessikat [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 02:07 | | Jessikat [Jessikat@Nightstar-2tikn6.dab.02.net] has joined #code |
| 02:09 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 02:15 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 02:15 | | mode/#code [+o macdjord] by ChanServ |
| 03:16 | | madgoat [gk.1wm.su@Nightstar-vugk4l.anchorfree.com] has joined #code |
| 03:17 | | madgoat [gk.1wm.su@Nightstar-vugk4l.anchorfree.com] has quit [RecvQ exceeded] |
| 04:27 | | gnolam [quassel@Nightstar-lv0k1q.cust-ip.bdtv.se] has quit [Ping timeout: 121 seconds] |
| 04:35 | | Turaiel[Offline] is now known as Turaiel |
| 04:37 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has quit [Ping timeout: 121 seconds] |
| 05:24 | | Derakon is now known as Derakon[AFK] |
| 05:43 | <&McMartin> | Except for the part where it's basically impossible to get acceptable images out afterwards, but I think I ultimately worked around that. |
| 05:53 | | Turaiel is now known as Turaiel[Offline] |
| 06:21 | | gnolam [quassel@Nightstar-lv0k1q.cust-ip.bdtv.se] has joined #code |
| 06:21 | | mode/#code [+o gnolam] by ChanServ |
| 06:22 | | celticminstrel [celticminst@Nightstar-epbo2s.dsl.bell.ca] has quit [Ping timeout: 121 seconds] |
| 06:22 | | celmin [celticminst@Nightstar-cue2qp.dsl.bell.ca] has joined #code |
| 06:22 | | mode/#code [+o celmin] by ChanServ |
| 06:30 | | Jessikat` [Jessikat@Nightstar-mt1sa9.dab.02.net] has joined #code |
| 06:31 | | gnolam_ [quassel@Nightstar-lv0k1q.cust-ip.bdtv.se] has joined #code |
| 06:33 | | Jessikat [Jessikat@Nightstar-2tikn6.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 06:35 | | gnolam [quassel@Nightstar-lv0k1q.cust-ip.bdtv.se] has quit [Ping timeout: 121 seconds] |
| 07:03 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:16 | | Jessikat` is now known as Jessikat |
| 07:28 | | gnolam_ [quassel@Nightstar-lv0k1q.cust-ip.bdtv.se] has quit [[NS] Quit: Gone] |
| 08:32 | | Kindamoody is now known as Kindamoody|afk |
| 08:39 | | macdjord is now known as macdjord|slep |
| 08:45 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [[NS] Quit: Subject 10: Developed dental hypertrophy, ocular luminescence, & extreme behavioural changes. Killed & partially consumed 2 graduate students. Euthanized with sustained automatic weapons fire.] |
| 09:02 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 09:02 | | mode/#code [+o macdjord] by ChanServ |
| 09:18 | | Jessikat [Jessikat@Nightstar-mt1sa9.dab.02.net] has quit [[NS] Quit: Bye] |
| 09:18 | | Jessikat [Jessikat@Nightstar-mt1sa9.dab.02.net] has joined #code |
| 09:18 | | Jessikat [Jessikat@Nightstar-mt1sa9.dab.02.net] has quit [[NS] Quit: Bye] |
| 10:42 | | Jessikat [Jessikat@Nightstar-mt1sa9.dab.02.net] has joined #code |
| 13:00 | | Jessikat` [Jessikat@Nightstar-injkum.dab.02.net] has joined #code |
| 13:04 | | Jessikat [Jessikat@Nightstar-mt1sa9.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 13:12 | | Jessikat` is now known as Jessikat |
| 13:16 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has quit [Server shutdown] |
| 13:16 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has quit [Server shutdown] |
| 13:16 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has quit [Server shutdown] |
| 13:16 | | Netsplit Deepthought.Nightstar.Net <-> Kakrafoon.Nightstar.Net quits: @iospace, @abudhabi, @Syloq, @Derakon[AFK], @Attilla, @Pi, @Vornicus, @Reiv, @Kindamoody|afk, @celmin, (+13 more, use /NETSPLIT to show all of them) |
| 13:17 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has joined #code |
| 13:17 | | Netsplit over, joins: @Reiv, Jessikat, @Attilla, @Kindamoody|afk, @PinkFreud, &ToxicFrog, Azash, ~Vornicus, Kizor, @abudhabi (+13 more) |
| 13:17 | | ServerMode/#code [+oooooaoqooooaoaoaoaoaoooo crystalclaw Reiv Attilla Kindamoody|afk PinkFreud ToxicFrog ToxicFrog Vornicus Vornicus abudhabi celmin JustBob [R] [R] starkruzr starkruzr Derakon[AFK] Derakon[AFK] jerith jerith jeroud jeroud Pi Syloq iospace] by *.Nightstar.Net |
| 13:26 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has joined #code |
| 13:27 | | mode/#code [+o Alek] by ChanServ |
| 13:53 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 14:08 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has joined #code |
| 14:08 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 16:34 | <@abudhabi> | I wonder: How lucrative would it be for someone who isn't at risk of death from old age to learn COBOL? |
| 16:41 | < Kizor> | No, don't do it, you have too much to live for, etc. |
| 16:41 | <@abudhabi> | Nah, I don't intend to do it myself. |
| 16:41 | <@abudhabi> | My preferred source of income is investments. |
| 16:42 | <@abudhabi> | But hypothetically, what if someone were autist enough? |
| 16:57 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has quit [[NS] Quit: Leaving] |
| 18:34 | <&ToxicFrog> | "if someone were autist enough"? Really? |
| 18:34 | <&ToxicFrog> | To answer your actual question, my understanding is that if you are fluent in COBOL and can get a contract with one of the big banks, you can retire young if you don't spend it all on whiskey. |
| 18:34 | <&ToxicFrog> | But this is based on hearsay, not hard numbers. |
| 18:50 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 19:00 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 19:00 | | Jessikat [Jessikat@Nightstar-injkum.dab.02.net] has quit [The TLS connection was non-properly terminated.] |
| 19:00 | | Jessikat [Jessikat@Nightstar-injkum.dab.02.net] has joined #code |
| 19:03 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has quit [Ping timeout: 121 seconds] |
| 19:07 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has joined #code |
| 19:07 | | mode/#code [+o Alek] by ChanServ |
| 19:10 | | Jessikat` [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 19:30 | | Jessikat`` [Jessikat@Nightstar-822ma9.dab.02.net] has joined #code |
| 19:32 | | Jessikat [Jessikat@Nightstar-injkum.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 19:43 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
| 19:43 | | mode/#code [+ao ToxicFrog ToxicFrog] by ChanServ |
| 19:57 | <&McMartin> | COBOL is not a difficult language |
| 19:58 | <&McMartin> | It is problematic in the modern era because it is a domain-specific language for operating on streams of preformatted records, and we've been in a bytestream world for many decades |
| 20:02 | < RchrdB> | Does COBOL not have the old Pascal problem wherein you can't (easily) operate with arbitrary-sized data structures in memory because everything is fixed-size at declared up front? |
| 20:05 | <&McMartin> | "domain-specific language for operating on streams of preformated records" |
| 20:05 | <&McMartin> | There's no such *thing* as a truly arbitrary-sized data structure, because punch card records are not arbitrary-sized |
| 20:06 | | Jessikat`` [Jessikat@Nightstar-822ma9.dab.02.net] has quit [[NS] Quit: Bye] |
| 20:06 | <&McMartin> | The thing is, that problem is hardly fatal and if you're doing *maintenance* as opposed to forward-design that's very firmly Not Your Problem. |
| 20:07 | <@Tamber> | But as long as you can bludgeon your byte-streams into appearing to be punch card records, COBOL can be made to do a quite scary amount of stuff. |
| 20:07 | <&McMartin> | It turns out it was *quite good at* those things, yes. |
| 20:08 | <&McMartin> | For all everyone rails at COBOL, they also hail the name most associated with its creation as one of the greats of our field. |
| 20:09 | <&McMartin> | Also it seems dynamic capacity tables were added in COBOL 2014~ |
| 20:10 | <&McMartin> | "In 2006 and 2012, Computerworld surveys found that over 60% of organizations used COBOL (more than C++ and Visual Basic .NET) and that for half of those, COBOL was used for the majority of their internal software.[7][100] 36% of managers said they planned to migrate from COBOL, and 25% said they would like to if it was cheaper. Instead, some businesses have migrated their systems from expensive mainframes |
| 20:10 | <&McMartin> | to cheaper, more modern systems, while maintaining their COBOL programs.[7]" |
| 20:10 | <&McMartin> | The thing is, I kind of remember back during the early micro era |
| 20:10 | <&McMartin> | Lots of much smaller shops would have some micro that had a bunch of custom BASIC code that did the things they and only they needed |
| 20:10 | <&McMartin> | Usually slammed together by the proprietor over a series of lazy weekends |
| 20:11 | <&McMartin> | And we're talking 10 PRINT "OMGWTFBBQ":GOTO 10 here, not VB |
| 20:11 | <&McMartin> | I have to wonder how much of the COBOL still out there is basically the businesses-big-enough-to-need-a-mainframe equivalent of that. |
| 20:24 | <&McMartin> | ... OK, and flipping through the Wiki page, I see one old-school feature in COBOL that (a) has stuck around and (b) seems of genuine value and (c) I have never seen copied |
| 20:25 | <&McMartin> | Which is comments that imply a form feed |
| 20:46 | < Jessikat`> | Hm? |
| 20:49 | < RchrdB> | It's perfectly consistent to have enormous respect for Rear Admiral Grace Hopper while at the same time very strongly desiring to not use COBOL. |
| 20:49 | <&McMartin> | "When printing this source code file, this comment begins a new page" |
| 20:50 | <&McMartin> | RchrdB: There's a difference between "strongly desiring no to use COBOL" and "the dread name must not be spoken without the Gestures Of Ritual Banishment", and the Internet is way more on the latter side of that scale |
| 20:50 | <&McMartin> | It is admittedly less hilarious with COBOL than with Ada |
| 20:50 | <&McMartin> | Because COBOL, like PHP, is a weird mutant offshoot language designed by people wholly outside of the core computing community |
| 20:51 | <&McMartin> | While Ada, at the end of the day, basically Won The War |
| 20:51 | < RchrdB> | I don't even know what you mean by that. |
| 20:51 | <&McMartin> | Ada was widely derided at the time of its creation, release, and update for being verbose and clunky and wholly unsuitable for any real projects |
| 20:52 | < RchrdB> | COBOL is still being used in mainframes that process $scary USD/hour. |
| 20:52 | <&McMartin> | The specific points of complaint about them are now the things that primarily define our basic expectations of OO languages |
| 20:52 | <&McMartin> | Oh, that part |
| 20:52 | < RchrdB> | Ada is being used for.... not as much software as COBOL is? |
| 20:52 | <&McMartin> | Yes, and PHP powers like 90% of the dynamic web |
| 20:52 | <&McMartin> | Right, Ada isn't being used for as much stuff |
| 20:52 | < RchrdB> | It's specifically the "Won The War" thing that I didn't understand |
| 20:53 | <&McMartin> | Because all those heinous philosophical objections underpin Java and C# as well, to the degree that flashing a sample of any of the three on a slide for a half a second and I don't think I could beat chance in identifying the language it was in. |
| 20:53 | < RchrdB> | aha fair enough |
| 20:53 | < RchrdB> | well |
| 20:53 | <&McMartin> | If I *know it's coming* I can check for keywords |
| 20:53 | < RchrdB> | You've got to read those early critiques bearing in mind the state of Ada implementations at the time? |
| 20:53 | <&McMartin> | Kiinda |
| 20:54 | <&McMartin> | "Fully qualified names everywhere by default" is, well |
| 20:54 | <&McMartin> | It turns out namespacing is really important you guys |
| 20:54 | < RchrdB> | no no I don't mean that one |
| 20:54 | <&McMartin> | Rendezvous was not the ultimate solution either |
| 20:54 | <&McMartin> | I'm not used to implementation level objections, bth |
| 20:54 | <&McMartin> | *tbh |
| 20:54 | < RchrdB> | I mean there were legit critiques like "the existing Ada compilers are buggier than the existing C compilers" |
| 20:54 | < RchrdB> | which was even true! briefly! |
| 20:55 | < RchrdB> | A modern analogue is that old jwz rant which I can't find right now about why Netscape didn't use C++ |
| 20:55 | <&McMartin> | OK, yeah, I'm excluding those |
| 20:55 | <&McMartin> | Ha ha ha |
| 20:55 | <&McMartin> | I don't need the link, fuckin' lived i |
| 20:55 | <&McMartin> | t |
| 20:55 | < RchrdB> | "because C++ compilers are buggy and broken and templates are unreliable" kind of thing |
| 20:55 | <&McMartin> | C++ changed incompatibly at least three times in my lifetime with it |
| 20:55 | < RchrdB> | except he wrote another one much more recently |
| 20:55 | < RchrdB> | because people kept quoting it out of context |
| 20:56 | <&McMartin> | Yeah, livin' that too |
| 20:56 | < RchrdB> | the second rant amounts to "boys and girls, I wrote that in like 1992. the implementation quality issues were true in 1992. that doesn't mean they're still true. g++ is lovely" |
| 20:56 | <&McMartin> | My TL;DR for that is "man, I'm glad I messed around with Rust a bit, otherwise I'd be a much worse C++11 programmer" |
| 20:56 | <&McMartin> | g++ didn't become "lovely" until egcs usurped it, imo, which I think was somewhere around when HL2 came out |
| 20:57 | < RchrdB> | the second rant I'm alluding to was, I think, written some time post 2010 |
| 20:57 | <&McMartin> | And, well, no, not even then. C++ is a fundamentally unlovely language, bu there's a whole lot of niches where it is the only quality option |
| 20:57 | < RchrdB> | also "g++ is lovely" wasn't a thing that jwz actually wrote :) |
| 20:58 | < Jessikat`> | learning Scheme is what made me a good C++ programmer |
| 20:58 | < RchrdB> | what he actually said was something more along the lines of "yes I know that free C++ compilers now exist in which the C++98 object model basically works as specified and templates actually work and templates are actually reliable now STFU and stop quoting the thing I wrote back in 1992-ish" |
| 20:59 | <&McMartin> | I was raised by a mathematician who went into the software industry, so I encountered Scheme far earlier than is traditional :) |
| 20:59 | <&McMartin> | This gave me a leg up when it was the language my university's first CS class was taught in |
| 20:59 | < Jessikat`> | I got it in 2nd year of uni |
| 20:59 | < Jessikat`> | I think that course had the single biggest effect of anything I've learned bar the evolutionary algorithms one |
| 21:05 | | * Vornicus needs to learn scheme too |
| 21:05 | <~Vornicus> | my first experience with a lispoid was common lisp itself |
| 21:06 | <~Vornicus> | Which pissed me off mostly because I already knew about libraries and it felt like... |
| 21:07 | < Jessikat`> | Clojure is truly worth learning IMO |
| 21:07 | < Jessikat`> | Scheme is great from an academic perspective |
| 21:07 | < Jessikat`> | but Clojure is truly the bleeding edge of a lot of concepts IMO |
| 21:08 | <&McMartin> | That sounds about right |
| 21:08 | <&McMartin> | I haven't gotten close enough to that edge in my experiments with it to opine properly |
| 21:09 | < Jessikat`> | in terms of value semantics, algorithmic specification via sequences and reducers and concurrency, I think it's well worth learning |
| 21:09 | < Jessikat`> | Or just watch Rich Hickey's talks on things |
| 21:10 | <&McMartin> | Yeah, it's the reducers/transducers stuff where I'd like to learn more, but this requires spare brain that I have not had in months |
| 21:11 | | Jessikat` is now known as Jessikat |
| 21:13 | < Jessikat> | ah, transducer's the word I was after |
| 21:22 | | Kindamoody|afk is now known as Kindamoody |
| 21:38 | | Jessikat [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 21:41 | | Kindamoody is now known as Kindamoody[zZz] |
| 21:50 | <@ErikMesoy> | "When you see a five digit Unicode number you know you're in trouble." |
| 22:00 | < RchrdB> | U+1F638 GRINNING CAT FACE WITH SMILING EYES ð¸ |
| 22:01 | < RchrdB> | One of the lovely side-effects of emoji being popularised is lots of software getting market pressure to fix bugs in their handling of Unicode, particularly with astral code points. |
| 22:12 | | Jessikat [Jessica@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 22:32 | <&McMartin> | Quite |
| 22:32 | | Syloq [Syloq@NetworkAdministrator.Nightstar.Net] has quit [[NS] Quit: .] |
| 22:32 | <&McMartin> | My favorite is still U+1F574 MAN IN BUSINESS SUIT LEVITATING ð´ |
| 22:33 | < RchrdB> | say waaaaat |
| 22:33 | <&McMartin> | All the more inexplicable for having originally been a corporate logo |
| 22:33 | < RchrdB> | ahhh |
| 22:33 | < RchrdB> | You know about the round-trip thing, right? |
| 22:33 | | Syloq [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 22:33 | | mode/#code [+o Syloq] by ChanServ |
| 22:33 | <&McMartin> | Uh, that seems like it could cover a lot of things. Unpack that? |
| 22:34 | < RchrdB> | Unicode's mission requires them to have a code point corresponding to every distinct code point in every single humanly-used (numberâcharacter) mapping in the world other than Unicode. |
| 22:34 | < RchrdB> | The round-trip property is that "convert-from-unicode . convert-to-unicode" is required to be the identity function for any script |
| 22:35 | <&McMartin> | Okay |
| 22:35 | < RchrdB> | so Unicode's mission requires it to be a (non-strict) superset of e.g. every single ISO-8859-x codepage, all the characters in shift-JIS, etc |
| 22:36 | < RchrdB> | and apparently one of the reasons for a lot of the *early* emoji being in Unicode is that a bunch of east Asian companies doing stuff with mobile phones went through a stage of making up their own character sets so that people could send each other cute pictures on their phones |
| 22:37 | <&McMartin> | I can attest that this is still very much a thing though now one sends references to pre-approved PNGs instead ("stickers"). |
| 22:37 | < RchrdB> | Yes. Apparently back in the day lots of people did this by making up their own code-pages and semi-proprietary-ish extensions of shift-JIS and friends. |
| 22:38 | <&McMartin> | Makes sense |
| 22:38 | < RchrdB> | so they could send those pictures in, say, two or three bytes, rather than <img src="picture-of-a-face.jpg"> which is like 20 bytes |
| 22:38 | | * McMartin gestures at code points 1 and 2 in code page 437 |
| 22:38 | < RchrdB> | it was the 90s! xml had not won yet! :) |
| 22:39 | < RchrdB> | and AFAIK this is why unicode contains code points for things like the Japanese post office logo |
| 22:39 | < RchrdB> | no other country's post office got their logo put into a character set that mobile messenger companies implemented but the Japanese post office did |
| 22:42 | <&McMartin> | While I suppose it is *possible* this is how MAN IN BUSINESS SUIT LEVITATING was added, it seems implausible since ISTR said logo was a smallish US record company |
| 22:43 | <&McMartin> | (Note that a correct rendering of it, though, based on this, requires the man to be standing at attention while levitaing, so that he and his shadow form something vaguely like an exclamation point. Many modern incarnations have him hovering Dhalsim-style in a Lotus position and this is Clearly Wrong.) |
| 22:44 | <&McMartin> | AHA |
| 22:44 | <&McMartin> | It was in Webdings |
| 22:44 | <&McMartin> | So There You Go, I guess |
| 22:45 | < RchrdB> | I bet the existence of Webdings and Wingdings must have made the Unicode consortium members so angry |
| 22:45 | < RchrdB> | â¡How DARE you use LATIN CAPITAL LETTER J for a fucking smiley face you invertebrate bastards!â |
| 22:45 | <&McMartin> | Dingbats have been a thing for hundreds of years |
| 22:46 | <&McMartin> | Smiley faces in the core 8-bit encoding go back to 1981 at the absolute latest |
| 22:46 | <&McMartin> | (the aforementioned Code Page 437) |
| 22:46 | <&McMartin> | http://www.newsweek.com/2016/05/06/secret-ska-history-man-business-suit-levitati ng-emoji-442192.html |
| 22:48 | < RchrdB> | to be clear, the sin here is not "using some code points to encode some dingbats" it's "overloading the meaning of numbers in the range 0x20 through 0x7f to mean completely different things than the corresponding ASCII points, via an out-of-band font switch" |
| 22:48 | < RchrdB> | making up your own code page with pictures and smilies in the 0x80 to 0xff range is more or less legit by comparison |
| 22:50 | <&McMartin> | Ehn. This requires us to also excoriate the Symbol font |
| 22:50 | <&McMartin> | (Surprisingly few Cyrillic encodings though!) |
| 22:50 | <&McMartin> | ASCII didn't become the universal subset until... kinda late, I think. |
| 22:53 | < RchrdB> | All the DOS codepages / ISO-8559-* left ASCII alone. |
| 22:53 | < RchrdB> | AFAIK at least some of the pre-Unicode standards for Chinese, Japanese and Korean more or less left ASCII alone too? |
| 22:53 | <&McMartin> | Yes |
| 22:54 | <&McMartin> | As did about half the Russian ones |
| 22:54 | <&McMartin> | Gotta be able to properly render English BASIC code |
| 22:54 | <&McMartin> | But, like |
| 22:54 | < RchrdB> | Like I think Shift-JIS had in-band bank-switching things, and the state machine initially started out in an ASCII-compatible page |
| 22:54 | <&McMartin> | Neither the C64 nor the early Sinclair machines match ASCII even if you restrict to 0x20-0x7f |
| 22:55 | <&McMartin> | The CJK stuff really pioneered the multibyte encodings |
| 22:55 | <&McMartin> | But yes, beyond DOS, EBCDIC took entirely too long to die |
| 22:56 | <&McMartin> | C64 has bankswiching as well but if you bankswitch in mixed case, upper and lower case characters are actually reversed |
| 22:56 | < RchrdB> | Did you hear about the IBM EBCDIC C++ thing? |
| 22:56 | <&McMartin> | I did not |
| 22:56 | < RchrdB> | There was a big controversy. C++11 or C++14, I can't remember which, wanted to drop EBCDIC support, I think including trigrams. |
| 22:57 | < RchrdB> | IBM had a loud public trouser-shitting moment because IBM is still selling "mainframe that you write programs for in C++ in EBCDIC" to customers and expects to for the foreseeable future. |
| 22:58 | <&McMartin> | Ah yes. |
| 22:58 | < RchrdB> | IBM kind of lost that argument though because they are the only remaining users of EBCDIC and no other C++ users were happy to hamstring their own efforts just to help out IBM. |
| 22:58 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 22:58 | < RchrdB> | So the dropping of whatever specific thing it was that was necessary for EBCDIC went ahead |
| 22:59 | < RchrdB> | and IBM are presumably going to have to sell "almost C++14 but working in EBCDIC" instead of being able to sell "C++14 in EBCDIC" I think. |
| 22:59 | < RchrdB> | You've got me curious now, how different were the C64 and Sinclair's character sets? |
| 22:59 | <&McMartin> | That would, I think, ultimately be a "C++14-with-exttensions" |
| 23:00 | < RchrdB> | Like, did they depart from the core codepoints like 0x30 == '0', 0x41 == 'A' and 0x61 == 'Z'? |
| 23:00 | <&McMartin> | Yes and no |
| 23:01 | <&McMartin> | C64 replaced the lower case letters with graphical characters and the control characters are wildly different from anything else, allowing fairly powerful control of the console with just strings and with no escape sequence necessary |
| 23:01 | <&McMartin> | The entire 0x80+ area is copies of lower level stuff or more control characters, probably due to don't-cares in the character decoder |
| 23:01 | < RchrdB> | I'm cool with that. 0x1f and below was always a minefield of incompatible shit anyway. |
| 23:02 | <&McMartin> | Then it had a mixed-case mode, which is very close to ASCII but the upper and lowercase letters are reversed |
| 23:02 | <&McMartin> | I have always assumed this was so it would be obvious at a glance which codepage you were using ("is the prompt in lower case?") |
| 23:02 | < RchrdB> | ...that's a little obscene. :) |
| 23:02 | <&McMartin> | Sinclair: Spectrum and later are ASCII |
| 23:03 | <&McMartin> | Pre-Spectrum Sinclair computers have a character encoding with no relation to any other system, tied, IIRC, to the details of its display logic. |
| 23:03 | <&McMartin> | It is apparently Actually Relevant, for the ZX81, that the newline character is the same codepoint as the opcode for the Z80's HALT instruction. |
| 23:04 | < RchrdB> | http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n4210.pdf |
| 23:04 | < RchrdB> | Might even have been C++17 in which trigrams go away |
| 23:05 | <&McMartin> | (The ZX81 seems to draw lines of text by setting up an interrupt handler that causes it to *jump to the video matrix*. Intervening circuitry notices this and feeds the CPU a series of no-ops until a HALT statement is reached.) |
| 23:05 | < RchrdB> | at some point I kind of have to ask pointed questions like "can't you chucklefucks just compile your programs on a real computer" |
| 23:06 | <&McMartin> | (The chip that intervened has the spectacular name of "Uncommitted Logic Array", which sounds like it can only be barely bothered to do its job but in fact is one of several standard predecessor technologies to the FPGA") |
| 23:06 | < RchrdB> | or "can't you chucklefucks just have your source code in ASCII even though everything else on the computer is EBCDIC" |
| 23:06 | <&McMartin> | "Sure, right after we convert every text file in our entire datacenter" |
| 23:07 | < RchrdB> | or "can't you chucklefucks just *write* your code on a real computer, copy it onto the idiot mainframe still in ASCII format, compile it there and just don't TELL anyone that the bytes aren't in EBCDIC" |
| 23:07 | <&McMartin> | That question is almost exactly equivalent to "can't you chucklefucks just run Python 3 for everything" |
| 23:07 | < RchrdB> | McMartin, no. I am not suggesting anything like that. |
| 23:08 | < RchrdB> | I am suggesting: 1. even an EBCDIC computer can have binary files on it. 2. treat the ASCII source code as though it's some weird binary file, 3. run a (remote) text editor on some other machine. |
| 23:08 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has quit [Connection closed] |
| 23:12 | <&McMartin> | Oh, right, also, because I forgot |
| 23:12 | <&McMartin> | ASCII itself changed over the years, so PETSCII also does not have ^ or _ |
| 23:12 | <&McMartin> | Instead having the older "Up arrow" and "Left Arrow" in those codepoints. |
| 23:13 | | Alek [Alek@Nightstar-7or629.il.comcast.net] has joined #code |
| 23:13 | | mode/#code [+o Alek] by ChanServ |
| 23:13 | <&McMartin> | Also, it looks like PETSCII in fact will not roundtrip: https://en.wikipedia.org/wiki/PETSCII |
| 23:13 | <&McMartin> | It looks like it's missing some fraction of its boxdrawing characters. |
| 23:15 | <&McMartin> | ... And now that I check, the Spectrum replaces ^ with up-arrow, but keeps _ and replaces ` with £ |
| 23:15 | < RchrdB> | Presumably the Unicode committee skipped 'em because they couldn't find anyone using PETSCII. |
| 23:16 | <&McMartin> | I don't begrudge stealing a codepoint for £ for an early UK-specific computer |
| 23:16 | < RchrdB> | Pretty good tagline. Commodore PET: deader than Linear-B! |
| 23:16 | < RchrdB> | er |
| 23:16 | < RchrdB> | Commodore PET: deader than Linear-B, according to the Unicode consortium! :D |
| 23:17 | <&McMartin> | PETSCII-art is still a live art form! |
| 23:17 | <&McMartin> | Here's the ZX81 character set, idly: https://en.wikipedia.org/wiki/ZX81_character_set |
| 23:17 | <&McMartin> | It does have the unusual privilege of being much easier to do convert-int-to-hex operations with than ASCII, because the digits *immediately* precede the letters. |
| 23:18 | <&McMartin> | It's also its own BASIC tokenization library, but it shared that propery with many other systems (albeit with different tokens). |
| 23:20 | < RchrdB> | Yeah putting '9' just before 'A' is a cute trick. |
| 23:22 | <&McMartin> | Man. |
| 23:25 | <&McMartin> | Speccy and Apple're basically the only 8-bit with curly braces. |
| 23:26 | <&McMartin> | ... and the TRS-80, I guess |
| 23:26 | <&McMartin> | I keep forgetting the TRS-80's an 8-bit. |
| --- Log closed Tue Apr 04 00:00:40 2017 |