00:04:26 -!- LKoen has joined.
00:04:28 -!- LKoen has quit (Client Quit).
00:11:02 -!- tromp has joined.
00:15:19 -!- tromp has quit (Ping timeout: 256 seconds).
00:29:05 -!- hppavilion1 has joined.
00:34:08 -!- tromp has joined.
00:35:06 -!- nooga has quit (Ping timeout: 240 seconds).
00:50:40 -!- ais523 has quit.
01:04:20 -!- tromp has quit (Remote host closed the connection).
01:10:16 -!- Lord_of_Life has quit (Excess Flood).
01:10:31 -!- Lord_of_Life has joined.
01:16:56 <fizzie> oerjan: FWIW, often there's a switch of some sort.
01:20:00 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
01:27:28 <oerjan> fizzie: the fridge is adjustable, but there's no switch in the freezer that i can find.
01:28:40 <oerjan> although there is a strange orange rope in the fridge which i'm not sure what does... it _might_ be connected to the freezer somehow.
01:29:25 -!- tromp has joined.
01:29:38 <oerjan> (they are not, however, a single box)
01:29:54 <shachaf> fizzie: Do you have an HTH kitchen?
01:38:08 -!- boily has joined.
01:38:13 <HackEgo> latin//LATINA EST SUBLIMISSIMA LINGUA MUNDI
01:48:04 <fizzie> shachaf: I'm not sure how that's defined.
01:48:27 <boily> hellochaf, VisuellozeR, fizziello.
01:50:52 <shachaf> Maybe the London real estate you buy can have an HTH kitchen.
01:53:45 -!- krok_ has quit (Quit: Leaving).
01:54:52 <fizzie> I looked a little at how those things work around here, and it seems really complicated.
01:55:20 <fizzie> At least compared to how we did it in Finland, where we just went to the bank we used anyway, asked for some money, and bought a place.
01:56:03 <shachaf> As opposed to what in London?
01:56:54 <fizzie> Apparently here the process involves at least a solicitor, a mortgage adviser, a lender, the seller, the seller's solicitor, the lender's surveyor, and probably some others as well.
01:57:54 <fizzie> And allegedly everyone remortgages their mortgages all the time.
01:58:19 <shachaf> Why do better deals keep showing up?
01:58:24 <fizzie> Well, every couple years.
01:58:43 -!- nooga has joined.
01:58:45 <shachaf> Are interest-only mortgage the best?
01:59:00 <doesthiswork> can you borrow money in finland for a place in london?
01:59:03 <fizzie> AIUI, you get a good deal for your first K years (for K of around 2 to 5), and then you remortgage when it would switch to the lender's default rate.
01:59:44 <fizzie> Plus there's freeholds, leaseholds, shared freeholds, commonholds, and I suspect a few other holds as well.
02:01:33 <fizzie> As far as I can tell, commonholds correspond to the most common arrangement (for flats) in Finland, where the people living in the building own a share in a company that owns the building, and then pays for someone to manage the things that need to be managed and makes the decisions that need to be decided.
02:02:02 <HackEgo> latin//LATINA EST SVBLIMISSIMA LINGVA MVNDI
02:02:13 <fizzie> But apparently no new-ish commonhold flats exist, because the developer companies can't extract as much money from the tenants that way.
02:03:24 -!- nooga has quit (Ping timeout: 260 seconds).
02:03:35 <shachaf> Sounds like a condominium thing in the US?
02:04:19 <fizzie> Maybe. All I know about condominiums is that the word is very fancy, and resembles CoDominium, which is a sci-fi universe.
02:05:38 <shachaf> Aha, Wikipedia confirms it.
02:06:11 <shachaf> The Suomi link points to Asunto#Asuinhuoneisto
02:06:52 <shachaf> fizzie: How should I get leverage?
02:07:02 <fizzie> I think that's just a general word for a flat.
02:07:21 <shachaf> Maybe real estate is the best way to do it.
02:07:32 <shachaf> I hear leverage is the best.
02:07:56 -!- tromp has quit (Remote host closed the connection).
02:08:01 <fizzie> The Asunto#Asuinhuoneisto section just says it's a thing meant for a place someone can live in.
02:12:50 <fizzie> I've been looking into setting up an alerting system, but I don't know how I could trust the alerting system to be working without something alerting me if it breaks.
02:14:17 <fizzie> I don't know yet. I want something that'd show up on things I'm looking at.
02:14:29 <fizzie> If you mean 'on what', I'd like to know when the wiki breaks, for example.
02:14:39 <fizzie> It's already got monitoring, but not alerting.
02:14:46 <shachaf> Not an alarm on your flat, for example.
02:14:50 <shachaf> What sort of monitoring has it got?
02:15:00 <shachaf> Shouldn't a monitoring system be able to alert automatically?
02:15:30 <fizzie> Three of the systems involved in the monitoring can do alerts, but I haven't configured any of them to do so.
02:16:57 <fizzie> There's a collectd instance on the wiki machine scraping stuff out of nginx (collectd can do alerts), it's sent to an InfluxDB for storage (the TICK stack includes a separate tool called Kapacitor for alerting), and I look at it with Grafana (the very latest Grafana versions do alerting).
02:17:41 <shachaf> Isn't Grafana more of a UI?
02:17:49 <fizzie> Yes, but they're building an alerting in it.
02:18:08 <shachaf> Of those systems it seems that the one I'd want to alert would be Kapacitor.
02:18:09 <fizzie> Apparently it was a heavily requested feature.
02:18:13 <shachaf> Well, I've never used Kapacitor.
02:18:38 <shachaf> collectd can presumably only alert on the values it collects, rather than queries based on historical values etc.?
02:18:59 <fizzie> Building alerting in Grafana may make sense so far as to it providing a nice user interface for configuring alerts.
02:19:49 <shachaf> I've been running Prometheus recently. It's a Borgmon clone.
02:20:10 <fizzie> So you said. I had a look at Prometheus, but I've spent a little bit too much time fiddling with the current setup already.
02:20:25 <fizzie> Incidentally, InfluxDB and the rest of TICK are also written in Go, as Prometheus.
02:20:36 <fizzie> https://zem.fi/tmp/esoqps.png
02:21:14 -!- tromp has joined.
02:24:31 -!- hppavilion1 has quit (Ping timeout: 256 seconds).
02:40:31 -!- hppavilion1 has joined.
02:41:36 -!- hppavilion2 has joined.
02:41:51 <shachaf> I think the data model is richer or something? I don't know.
02:43:30 -!- boily has quit (Quit: LIMERICK CHICKEN).
02:44:15 <shachaf> Did you see that Facebook released their time series database recently?
02:45:18 <shachaf> fizzie: that mean request processing time is p. high
02:45:19 -!- hppavilion1 has quit (Ping timeout: 245 seconds).
02:47:04 <shachaf> Everything is being written in Go nowadays. :-(
02:49:12 <fizzie> I think I heard something about the Facebook thing.
02:49:47 <fizzie> And Microsoft released something to do with large repositories in Git.
02:50:09 <fizzie> https://github.com/Microsoft/GVFS
02:51:02 <fizzie> "With GVFS, this means that they now have a Git experience that is much more manageable: clone now takes a few minutes instead of 12+ hours, checkout takes 30 seconds instead of 2-3 hours, and status takes 4-5 seconds instead of 10 minutes."
02:55:11 -!- myname has quit (Ping timeout: 240 seconds).
02:56:40 <fizzie> (I was going to say "huge" instead of "large", but turns out the repository they're talking about has just 3 million files.)
02:59:17 -!- myname has joined.
03:02:43 <shachaf> That's only a couple orders of magnitude smaller than Google's, according to http://cacm.acm.org/magazines/2016/7/204032-why-google-stores-billions-of-lines-of-code-in-a-single-repository/fulltext
03:03:03 <shachaf> Whenever I mention a Google thing (like CitC the other day), I feel obligated to include a citation.
03:16:36 <fizzie> I just look up a citation I *could* include, but don't.
03:17:27 -!- oerjan has quit (Quit: Nite [citation needed]).
03:22:13 -!- nal has quit (Quit: Going offline, see ya! (www.adiirc.com)).
03:24:40 -!- Sgeo has joined.
03:28:33 -!- VisualizeR has quit (Ping timeout: 258 seconds).
03:29:26 -!- Cale_ has joined.
03:29:30 -!- VisualizeR has joined.
03:36:13 -!- JX7P has quit (Ping timeout: 258 seconds).
03:37:23 -!- IRIXUser has joined.
03:37:46 -!- IRIXUser has changed nick to Guest7544.
03:40:50 -!- hppavilion2 has quit (Ping timeout: 240 seconds).
03:41:21 -!- Sgeo has quit (Read error: Connection reset by peer).
03:46:47 -!- nooga has joined.
03:52:06 -!- nooga has quit (Ping timeout: 276 seconds).
03:53:24 -!- hppavilion2 has joined.
04:00:11 -!- Zarutian has quit (Quit: Zarutian).
04:09:21 -!- MDude has quit (Read error: Connection reset by peer).
04:09:40 -!- MDude has joined.
04:15:25 -!- Cale has quit (Disconnected by services).
04:15:28 -!- Cale_ has changed nick to Cale.
04:15:48 -!- Cale_ has joined.
04:50:44 -!- hppavilion2 has quit (Ping timeout: 245 seconds).
04:52:41 -!- hppavilion2 has joined.
05:06:25 -!- Cale has quit (Read error: Connection reset by peer).
05:14:53 -!- tromp has quit (Remote host closed the connection).
05:23:04 -!- Cale has joined.
05:34:54 -!- nooga has joined.
05:37:55 <shachaf> What language should be used to specify build system configuration?
05:39:06 -!- nooga has quit (Ping timeout: 240 seconds).
05:44:01 -!- VisualizeR has quit (Ping timeout: 248 seconds).
05:47:48 <Cale> shachaf: At the company I work for, we use nix. I really wish that nix's language was just Haskell.
05:48:18 <shachaf> Nix's language seems very complicated.
05:48:30 <shachaf> But also it doesn't really seem like a build system configuration language?
05:48:39 <shachaf> It doesn't replace Makefiles/cabal/whatever.
05:48:53 <Cale> It's not so much that the language itself is complicated, it's that the way people use it is complicated and often undocumented, and impossible to discover because it's untyped.
05:49:25 <Cale> Well, cabal2nix is involved
05:50:11 <shachaf> Bazel and all Bazel clones use Python. It's not ideal but it's not bad.
05:50:15 <shachaf> I'm not sure what would be better.
05:52:19 <shachaf> Would Haskell be better? Unclear.
05:53:32 <Cale> If nix were simply built in Haskell rather than its own weird inane language, it would be really nice.
05:54:16 <shachaf> Nix seems to be at the wrong level of abstraction, though.
05:54:27 <shachaf> But maybe I just don't understand Nix.
05:54:51 <Cale> There needs to be some sort of transition project which starts out by implementing nix in Haskell with Haskell as the configuration language, and then building a nix interpreter which integrates with that so that all the existing work on .nix files can be leveraged and gradually replaced.
05:55:20 <shachaf> I went to a talk about Reflex FRP the other day, by the way.
05:55:49 <Cale> I've heard the name
05:56:00 <Cale> I'm in NYC for the next couple of weeks btw
05:56:26 <shachaf> NYC pretty good. I should go sometime.
05:56:50 <Cale> Then I'm going to Melbourne for a month off :)
05:57:19 <Cale> Ah, perhaps I should have... it's all already booked though.
05:57:38 <Cale> My stopover is in Hong Kong
05:57:55 <Cale> Gonna be a crazy flight :)
06:00:53 <shachaf> Cale: You should come to BayHac 2017 in Apr.
06:02:11 <Cale> oh man... might be too soon after all this, but we'll see :)
06:04:14 <Cale> Yeah, it's probably not really all that hard -- I just fully expect to be fairly burnt out on travel by the time I get back home toward the end of March.
06:04:53 <Cale> I'm only going to be home for two days between now and March 20
06:06:06 <shachaf> So you have several weeks to rest before BayHac
06:06:20 <shachaf> I'll probably be in NYC in Apr.
06:06:41 <shachaf> But it shouldn't conflict, so it'll all be fine.
06:58:50 -!- PinealGlandOptic has joined.
07:07:21 -!- tromp has joined.
07:12:02 -!- tromp has quit (Ping timeout: 258 seconds).
07:14:54 -!- Cale has quit (Ping timeout: 276 seconds).
07:23:07 -!- nooga has joined.
07:27:39 -!- nooga has quit (Ping timeout: 252 seconds).
08:30:33 -!- dos has joined.
08:33:39 -!- hppavilion2 has quit (Ping timeout: 245 seconds).
08:46:53 -!- doesthiswork has quit (Quit: Leaving.).
09:11:04 -!- nooga has joined.
09:15:36 -!- nooga has quit (Ping timeout: 240 seconds).
09:16:22 -!- tromp has joined.
09:20:36 -!- tromp has quit (Ping timeout: 240 seconds).
09:27:38 -!- mroman has joined.
09:28:08 <mroman> I fear the day captchas prove to me that I'm a robot.
09:28:27 <mroman> like when I click on "I'm not a robot" and it doesn't trust me.
09:28:45 -!- nooga has joined.
09:28:57 <mroman> there's this funny video of a guy beating the "I'm not a robot captcha" with a robot.
09:30:59 -!- AnotherTest has joined.
09:39:49 <mroman> stupid youtube video editor is so fucking buggy
09:47:10 -!- nooga has quit (Ping timeout: 240 seconds).
10:03:10 -!- nooga has joined.
10:05:43 <mroman> I'm making very quick guides on how to do stuff
10:05:52 <mroman> following the ms paint quality of graphics tradition
10:05:57 <mroman> mostly because I suck at graphics :)
10:22:00 <int-e> . o O ( The "MS Paint" school of art was a minimalist art movement in the late 20th and early 21st century. Its proponents were convinced that "name defines function" and therefore, the program called "Paint" would be a suitable medium for producing paintings. )
10:31:21 <mroman> name defines function .
10:32:40 <int-e> however, "Word" is apparently used to write books.
10:34:40 <mroman> I taught an SQL course last year
10:35:29 <mroman> I don't like coding for 8h a day
10:51:31 -!- nooga has quit (Quit: Lost terminal).
11:01:21 -!- nooga has joined.
11:18:02 -!- tromp has joined.
11:22:49 -!- tromp has quit (Ping timeout: 260 seconds).
11:26:11 -!- kiki` has quit (Quit: Connection closed for inactivity).
11:28:00 -!- VisualizeR has joined.
11:40:14 -!- dos has changed nick to hppavilion[1].
11:48:07 <hppavilion[1]> myname: I'm trying to translate "Grab 'em by the pussy" [for... fun?], specifically discussing whether it'd be "Greift" or "Greifst" or "Greife"
11:50:43 <myname> itjs imperative so either greif oder greift, depending on wether you talk to a single person or a crowd
11:51:05 <shachaf> but maybe you can do that somewhere else or nowhere at all
11:51:25 <shachaf> and in particular not here
12:01:56 -!- LKoen has joined.
12:04:51 -!- VisualizeR has quit (Ping timeout: 252 seconds).
12:14:15 -!- DHeadshot has joined.
12:18:34 -!- tromp has joined.
12:23:00 -!- tromp has quit (Ping timeout: 252 seconds).
12:38:48 -!- hppavilion[1] has quit (Ping timeout: 255 seconds).
13:23:54 <mroman> @msg hppavilion[1] "Packt sie an der Muschi"
13:24:15 <mroman> @tell hppavilion[1] "Packt sie an der Muschi"
13:25:46 <mroman> (but please don't ever use that)
13:27:24 <fizzie> fungot maintenance in progress, please stand by.
13:27:27 -!- fungot has quit (Remote host closed the connection).
13:28:05 -!- fungot has joined.
13:28:31 <fizzie> fungot: Are you feeling quite okay?
13:29:36 <fizzie> I see the incoming messages in the log, but it doesn't seem to be replying.
13:30:07 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf/ul, code=text/str:N; ^str 0-9 get/set/add [text]; ^style [style]; ^bool
13:30:30 <fizzie> That side works, why doesn't the babbling.
13:32:48 -!- fungot has quit (Remote host closed the connection).
13:34:07 -!- fungot has joined.
13:34:09 -!- fungot has quit (Remote host closed the connection).
13:35:20 <fizzie> Also that's not right, the proxy quits when fungot disconnects.
13:35:25 <fizzie> 05-02-2017 13:37:20 FATAL: Failed assertion in src/log.c(205): olf
13:35:50 <fizzie> Apparently you can turn logging off, but then it will crash whenever a client disconnects.
13:37:54 -!- fungot has joined.
13:38:19 <fungot> That Pong alone cannot stop!
13:38:33 <fizzie> Hmm. I must've messed something to do with saving the bot's own nickname.
13:40:41 <fizzie> Heh, I think it thinks its nickname is the password.
13:42:23 -!- fungot has quit (Remote host closed the connection).
13:42:31 <fizzie> ...still crashes when disconnecting.
13:43:55 -!- fungot has joined.
13:44:01 <fizzie> fungot: How about now?
13:44:02 <fungot> fizzie: it does that fnord to it ( programmer time saved, future extensibility, etc.
13:47:02 -!- Zarutian has joined.
13:47:38 <fizzie> @tell Phantom_Hoover You'll be glad to know fungot now supports a connection password and non-6667 ports, for convenient bouncer use.
13:47:38 <fungot> fizzie: it's like a type fnord ( or something). but when run from unix, and i don't know
14:11:28 <mroman> You know fungot is alive if he says fnord.
14:11:28 <fungot> mroman: it is. the fnord bus actually goes fnord. :p
14:19:37 -!- boily has joined.
14:33:05 <rdococ> where can I think with [the physical implications of] portals?
14:37:09 <rdococ> I'm bored and now I want to discuss the physics of portals... do you know a channel I can discuss such a thing without being met with silence?
14:38:47 <rdococ> ...just as I predicted, silence
14:43:34 -!- oerjan has joined.
14:48:09 <boily> sorry, don't know of any channel where.
14:48:24 <boily> (also, currently dieing at crawl.)
14:49:30 -!- gsora has quit (Quit: uh-oh!).
14:50:27 * oerjan wanted to make a wisdom where, but then remembered #esoteric is the only channel.
14:51:22 -!- gsora_ has joined.
14:53:29 <HackEgo> universal property//Universal properties are the best.
14:53:41 <HackEgo> high hat//A high hat is the same as a top hat, not the same as a hi-hat, just like how a top quark is not the same as an up quark.
14:54:21 <oerjan> or down. anyway, has its charm.
14:54:28 <oerjan> even if a bit strange.
14:56:17 <boily> physics are weird shit, yo.
14:57:14 <boily> but everything is made of the five elements: boom, orange, pickle, pungent and sweet.
14:57:35 <oerjan> is that discordianism or something
14:59:32 -!- tromp has joined.
15:04:25 -!- wob_jonas has joined.
15:09:30 -!- tromp has quit (Remote host closed the connection).
15:09:57 <wob_jonas> fizzie: why do you want to build an alert system? don't people on the channel already alert you?
15:10:59 -!- tromp has joined.
15:11:00 <wob_jonas> and this might be the first time I heard my wisdom cleans something up
15:11:22 <fizzie> I'm not always looking at the channel.
15:12:36 -!- VisualizeR has joined.
15:22:45 -!- LKoen has quit (Remote host closed the connection).
15:29:54 -!- myname has quit (Ping timeout: 245 seconds).
15:34:50 -!- myname has joined.
15:37:15 <HackEgo> [wiki] [[User:Arkenidar]] M https://esolangs.org/w/index.php?diff=50849&oldid=50710 * Arkenidar * (+79)
15:39:04 -!- S1 has joined.
15:43:05 <oerjan> fungot: feeling more bouncy today?
15:43:06 <fungot> oerjan: but the gui still needs to have the students who failed the exam take the supplementary."
15:43:55 <int-e> fungot: stop being coherent please
15:43:56 <fungot> int-e: although, if you watch movies on your computer? ( they'll actually come and check this!)
15:44:57 -!- tromp has quit (Remote host closed the connection).
15:46:25 <ffj-bot> wob_jonas: pnrmurnkbsttgffqsgqadmt]`]htjeqthsvpsvsnlgdrrfo
15:47:26 <oerjan> fungot: if you don't improve your incoherence ffj-bot will be beating you
15:47:27 <fungot> oerjan: i believe so, a mechanical amplifier... i have enough trouble keeping other /humans/ off my keyboard.
15:48:01 <ffj-bot> wob_jonas: wbebigxljdefqlzjyndvanxyommgniscsyqsgudffwfseti
15:49:25 -!- tromp has joined.
15:55:13 -!- ibrahemgebri has joined.
15:58:35 -!- ibrahemgebri has quit.
16:00:02 -!- DHeadshot_ has joined.
16:00:14 <mroman> let's turing test this fungot.
16:00:14 <fungot> mroman: i'm a bit short in the fnord before it totally befunged me.
16:00:33 -!- DHeadshot has quit (Ping timeout: 248 seconds).
16:00:47 -!- augur has joined.
16:07:31 -!- DHeadshot_ has changed nick to DHeadshot.
16:08:47 -!- idris-bot has quit (Ping timeout: 264 seconds).
16:10:15 -!- oerjan has quit (Quit: Later).
16:16:21 <boily> mrelloman. fungot is feeling a little short in the fnord. don't sentience him today.
16:16:21 <fungot> boily: ( the whole compiler uniformly used a cps lambda calculus for the people that i don't
16:17:00 <Zarutian> hmm... a rogue like where the floor is inscribed in befunge patterns and some enemies must follow the instructions.
16:22:14 -!- ais523 has joined.
16:22:57 -!- nooga has quit (Ping timeout: 248 seconds).
16:23:16 <boily> Zarutellon. more like battlebots trying to selfmodify the same program to trip the other ones into oblivion?
16:24:24 -!- tromp has quit (Remote host closed the connection).
16:31:01 -!- VisualizeR has quit (Quit: ( www.nnscript.com :: NoNameScript 4.21 :: www.esnation.com )).
16:33:37 -!- nooga has joined.
16:34:16 -!- gsora_ has changed nick to gsora.
16:35:11 <HackEgo> calzone//A calzone is a part of California. They include norcal and socal.
16:35:24 <boily> . o O ( wasn't it "nocal"? )
16:35:30 <HackEgo> ic//ic what you did there.
16:35:45 -!- gsora has quit (Quit: uh-oh!).
16:36:41 -!- gsora has joined.
16:36:48 -!- tromp has joined.
16:38:49 -!- Zarutian has quit (Quit: Zarutian).
16:41:31 <HackEgo> murphy'//Murphy's law obviously does not hold in wisdom/
16:44:29 -!- ais523 has quit (Remote host closed the connection).
16:45:39 -!- ais523 has joined.
16:46:32 -!- doesthiswork has joined.
16:52:54 -!- S1 has left ("Leaving").
17:01:35 -!- Zarutian has joined.
17:39:42 -!- LKoen has joined.
17:55:41 -!- boily has quit (Quit: ABOUT CHICKEN).
17:59:53 -!- tromp has quit (Remote host closed the connection).
18:05:26 -!- LKoen has quit (Remote host closed the connection).
18:10:30 -!- LKoen has joined.
18:16:47 -!- idris-bot has joined.
18:18:29 -!- LKoen has quit (Remote host closed the connection).
18:19:22 -!- ais523 has quit (Remote host closed the connection).
18:20:31 -!- ais523 has joined.
18:29:53 -!- DHeadshot has quit (Ping timeout: 248 seconds).
18:51:51 -!- idris-bot has quit (Ping timeout: 252 seconds).
19:00:44 -!- tromp has joined.
19:02:54 <wob_jonas> heh heh. I knew that Knuth's homepage has been changed: whereas it used to say that TAOCP vol 5 is estimated to be ready in 2020, now it says 2025. But I just realized that vol 3 second edition says that the reworked version of vol 3 is estimated at 2015. Funny.
19:05:20 -!- tromp has quit (Ping timeout: 255 seconds).
19:05:54 <fungot> mroman: this is one place where i used essentials of programming languages
19:08:23 -!- doesthiswork has quit (Quit: Leaving.).
19:10:30 -!- mroman has quit (Ping timeout: 260 seconds).
19:10:35 -!- mroman_ has joined.
19:17:03 <int-e> fungot: you seem to have mastered syntax, but what about semantics?
19:19:35 <wob_jonas> semantics? what does that mean? it's just a word
19:20:24 <int-e> wob_jonas: it was talking about essentials of programming languages
19:20:41 <int-e> wob_jonas: which for some inexplicable reason I associate with the string "syntax and semantics"
19:26:37 <ais523> "semantics" has a mathematical definition
19:27:03 <ais523> but in this channel, it's normally used to describe the method via which a programming language gives meaning to a parse tree
19:27:11 <ais523> (with the syntax being the way the source code is converted into a parse tree)
19:27:42 <wob_jonas> ais523: sure, I'm just being silly in a self-referential way, but it wasn't very funnt
19:28:11 <ais523> oh, I completely missed the joke, you need at least a pair of quotes for it to work though
19:30:15 <int-e> I saw the circularity but thought it was a troll attempt rather than a joke, sorry. (An honest and complete answer would become *very* lengthy, and I'm not sure I would be up to giving one.)
19:46:57 -!- doesthiswork has joined.
19:53:00 <HackEgo> Syntax is just a subset of grammar.
19:53:06 <HackEgo> Grammar is just the evil subset of syntax.
19:53:10 <HackEgo> Grammar is just the evil subset of syntax.
19:53:36 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: culprit: not found
19:57:32 -!- krok_ has joined.
20:11:04 -!- ais523 has quit (Remote host closed the connection).
20:12:14 -!- ais523 has joined.
20:17:15 -!- ais523 has quit (Remote host closed the connection).
20:18:25 -!- ais523 has joined.
20:23:32 -!- MoALTz has quit (Quit: Leaving).
20:46:31 -!- Guest7544 has quit (Changing host).
20:46:31 -!- Guest7544 has joined.
20:46:33 -!- Guest7544 has changed nick to JX7P.
20:47:15 <HackEgo> 1166) <fungot> int-e: such were the idle tales of the fnord
20:49:30 -!- doesthiswork has quit (Quit: Leaving.).
21:00:32 -!- tromp has joined.
21:05:05 -!- tromp has quit (Ping timeout: 248 seconds).
21:11:13 <wob_jonas> ‘Life,’ said Marvin, ‘don't talk to me about life.’
21:15:11 <zzo38> Do you know what is best way to do with OpenGL to replace a texture and a set of vertices with data from mmap()ed memory during every frame without lagging?
21:15:13 <int-e> Hmm, there was that message that made Marvin smile? "We apologize for the inconvenience.".
21:17:10 <Zarutian> zzo38: well there will be lag as the texture will have to be DMAed from main memory to graphics card memory
21:18:43 <zzo38> Zarutian: Yes, but I would want to try to reduce the amount of lag and also to reduce the amount of slowness
21:19:31 <zzo38> (And I want to copy a set of vertices as well as a texture, and then to render those things in a display list)
21:23:48 -!- hppavilion[1] has joined.
21:25:19 <zzo38> wob_jonas: I just wanted to know the best way to do; it does not have to be perfect. (Also, I haven't tried anything yet)
21:25:40 -!- hppavilion2 has joined.
21:27:09 <Zarutian> zzo38: the lag will always be there due to this bottleneck that is the bus between the graphics card and main memory. This is true even with systems that say that they use unified memory. (In that case, unless the memory is dual ported, it will cause contention between CPU and GPU. And often there is no mapswapping of memory banks. (CPU does stuff with memory bank A while GPU has memory bank B and then they switch)
21:28:02 <Zarutian> I think there is something called transfers sets or command lists but I havent looked at OpenGL (nor (in)DirectX for that matter) in a long while now.
21:28:08 <int-e> https://www.khronos.org/opengl/wiki/Buffer_Object ... no I'm not reading that now (and it won't give you the answers you want anyway... it'll depend on the hardware. for example, the hybrid CPU and GPU packages with shared memory have the benefit that no DMA is taking place. PCIe has reasonably fast access to main memory too, I believe.
21:29:20 -!- hppavilion[1] has quit (Ping timeout: 255 seconds).
21:29:31 <int-e> I think buffer objects are the modern abstraction for everything now, I mean look at this list of types: https://www.khronos.org/opengl/wiki/Buffer_Object#General_use
21:30:20 <wob_jonas> It's good to know there are people on this channel who know about gpu stuff
21:31:07 * int-e doesn't really... it's on my "I should take some time to properly (re)learn this" horizon)
21:32:34 <wob_jonas> it's on my list of things I don't want to learn about
21:36:44 <zzo38> Why do they make it so complicated? (Many older kind of computers do not make it so complicated and so confusing)
21:38:20 <Zarutian> int-e: re the hybrid CPU and GPU packages: see what I said about contention. (Which is more accuratly said to be port contention)
21:39:43 <int-e> Zarutian: yes, I know it's not a free lunch.
21:41:00 <Zarutian> zzo38: depends on which older computers. Many just give you a frame buffer memory that you can write pixels into.
21:41:31 <int-e> (In the end hybrid or standalone GPU will be just another variable for deciding where best to put your various buffer objects)
21:41:49 <Zarutian> int-e: I am just rather befuddled why they want to add pressure on the memory to GPU or CPU bottleneck.
21:43:52 <zzo38> My own design was just for CPU and video to be clock interleaved, and video memory can be mapped anywhere in the main memory space (usually RAM, but it does not have to be), and then a display list is executed during hblank in order to change the setting of what to draw on that scanline.
21:44:33 <Zarutian> zzo38: so basically, ATARI esque setup
21:44:58 <int-e> Zarutian: shared memory = simpler design? I've heard the point (I think in a talk by John Carmack) that a lot of data has to travel from the CPU/main memory to the GPU each frame *anyway* and having shared memory (= direct access of the GPU to that data) might very well make this cheaper overall... I'm skeptical but it's definitely not an obvious tradeoff.
21:45:00 <zzo38> Zarutian: Yes, somewhat, but more sophisticated.
21:45:55 -!- copumpkin has joined.
21:45:55 <int-e> Zarutian: I can see an unambiguous advantage in not sharing the frame buffer though.
21:46:01 <wob_jonas> actually, I think the video chipset being the most complicated part of the computer is the rule, ever since the first video terminals, and the IBM PC with its dumb simple video cards are the exception
21:46:41 -!- hppavilion2 has changed nick to hppavilion[1].
21:47:17 <Zarutian> int-e: perhaps, but I have seen old designs where they used crossbar switch between CPUs, GPUs, DMA drivers and memory banks.
21:48:39 <Zarutian> wob_jonas: the reason for that was the price of memory iirc versus logic die size. (If you could use a specialized logic to lessen memory consuption for video terminal stuff then you did)
21:49:48 <int-e> Zarutian: damn there was a cute story a while ago where somebody ran benchmarks and they got quite consistently faster after 30 minutes or so... that was with a shared framebuffer, and after 30 minutes the display would switch off, freeing a noticable fraction of the memory bandwidth (1920*1080*60*4 = 500MB/s, it would be more today). I have no idea how to find it again :-/
21:49:57 <wob_jonas> and I think part of the reason why memory capacity (as opposed to latency) became relatively cheaper compared to other stuff in the computer is partly that DRAM starts out big but scales better to larger RAM sizes
21:50:25 <Zarutian> int-e: the beuity of it was that there were usually no contention between diffrent accessors to same memory banks.
21:50:29 <int-e> wob_jonas: possible?
21:50:29 <wob_jonas> I did main memory throughput benchmarks
21:50:42 <wob_jonas> and found that it got faster when the screen saver started
21:50:58 <zzo38> Zarutian: But I think the Atari ANTIC only allows executing one display list instruction per hblank, and many settings could not be set by the display list. My own design does not have these limitations.
21:51:10 <wob_jonas> Zarutian: the contention was with the throughput. I didn't try to measure latency in that case.
21:51:53 <int-e> wob_jonas: did you just have graphs or also some text online? but I guess reading it here would be memorable enough for me.
21:51:55 <Zarutian> zzo38: sure. But the basic idea is the same.
21:52:37 <Zarutian> int-e: it just had one 'huge' drawback: cost. At least at that time.
21:52:42 -!- Phantom_Hoover has joined.
21:53:10 <Zarutian> (cost as in relatively expensive to manifacture than other cheaper yet worse methods)
21:53:12 <wob_jonas> If you had a complete copy of the logs, you could probably find it.
21:54:37 -!- hppavilion[1] has quit (Remote host closed the connection).
21:54:58 -!- hppavilion[1] has joined.
21:55:51 -!- PinealGlandOptic has quit (Quit: leaving).
21:56:04 <Zarutian> zzo38: I have somewhere in my ideas file an draft of an spec where instead of an application being only limited to framebuffers, textures, voxels, uniforms and simple gpu shader kernels, they could send a whole program expressed in a binary language that was on the level of primitive recursive function level and therefore guranteed to terminate.
21:57:11 <Zarutian> zzo38: a bit like ReGIS&Sixel, Ripscrip, NAPLPS or Apple QuickDraw commands list (PICT) in that regard
21:57:52 <int-e> Zarutian: probably still is expensive... otherwise more multicore CPUs would offer more than 2 memory interfaces.
21:58:18 <Zarutian> so feel free to discuss your ideas regarding this here, zzo38 ;-Þ
21:58:24 <int-e> (perhaps 4 is the norm now for high-end CPUs?)
21:59:49 <Zarutian> int-e: well that design didnt have multi core CPUs, just a number of CPUs (Called Computing Processing Units just to get the acronym to fit ;-)
21:59:58 <int-e> (just looking at intel and amd here, fwiw)
22:00:00 <zzo38> Zarutian: I would have had it a bit differently; the program is not guaranteed to terminate but if it does not terminate it will restart anyways during each vblank. That way loops and so on are possible and you can implement your own tiles perhaps
22:01:25 <wob_jonas> zzo38: is there much difference? A primitive recursive program could still easily loop as much that it terminates only after the universe has ended.
22:01:52 -!- tromp has joined.
22:02:24 <wob_jonas> As for CPU and memory interface, what I really wish for is for Intel to figure out a sane way in future x86_64 cpus how the OS can opt for a mode where all pages are at least 8K (or at least 16K) sized,
22:02:35 <int-e> Zarutian: sure, but to make it analogous, a multicore-CPU is just several CPU's, and possibly GPUs on a single die, communicating with external buses (memory, PCIe, more?) and in a perfect world there'd be a crossbar switch between all those components...
22:02:51 <zzo38> My own design did allow loops too and also did that (restarted the program during each vblank), but could not execute during rendering so you had to program the registers to control rendering during each scanline (it renders tiles eight pixels wide, or can render in pixel mode instead, but cannot mix them in a single scanline; also tiles are only one pixel high but the display list can change the character base address during each scanline in order
22:02:57 <wob_jonas> because the 4K page size is currently causing the cpus to not being able to have more than 8*4K L1 data cache
22:03:11 <wob_jonas> and that limitation would be so useful to get over.
22:03:13 -!- hppavilion[1] has quit (Ping timeout: 248 seconds).
22:03:17 <zzo38> Actually in my design, the display list program can't terminate, although there is a "wait for next scanline" instruction.
22:03:20 <Zarutian> zzo38: one trick I did way back when first there was a computer in the house and I was programming stuff in QBASIC was to keep a copy of video memory block where the mouse cursor was to be drawn to be used to undo the cursor drawing when the mouse was moved.
22:03:35 <int-e> Zarutian: the reality will be less than ideal and probably too complicated for my poor little brain :)
22:04:12 <Zarutian> zzo38: just because it otherwise took too long to redraw the screen and there wasnt enough memory to do dual buffering. Worked well for other sprite kind of stuff too.
22:04:50 <int-e> Zarutian: one cute point of the hybrid design is that a GPU can potentially access data by snooping CPU caches without going via main memory at all.
22:05:17 <wob_jonas> Zarutian: right, PC video cards didn't have built-in sprite support of any sort (and have other limitations too), which is why they're so hard to use.
22:05:43 <wob_jonas> People still ended up doing magical stuff with PC video cards of course, but still.
22:06:14 -!- tromp has quit (Ping timeout: 255 seconds).
22:06:20 <Zarutian> wob_jonas: I was rather pleased with my ten yearold self when I figured out that trick on my own.
22:06:41 <zzo38> My own design does include built-in sprite support though, but the display list program still has to alter the sprite Y scroll register during each scanline
22:07:42 <zzo38> (Also my own design has no display list interrupt like ANTIC has, because it is unnecessary. There is still the vblank interrupt though.)
22:08:14 <Zarutian> wob_jonas: and my brother was rather impressed that I had added mouse cursor (and gui buttons) to an fractal drawing program that we had copied from an photocopied then faxed computer programming magazine article.
22:08:46 <int-e> Zarutian: hmm, and all that sounds like I really like those hybrids, and I actually don't... but the tradeoffs are all but clear, and the reason is that both the hybrids and the standalone GPUs are really hard to reason about.
22:09:00 <wob_jonas> Zarutian: did that use machine code (non-BASIC) for some of the computation?
22:09:01 <int-e> (performance wise)
22:10:12 <Zarutian> wob_jonas: none what so ever, took a while to draw the fractal, just did a ?interrupt? check in the next innermost loop iirc
22:10:19 <zzo38> A program I wrote in QBASIC to implement a solitaire card game used a XOR picture for the mouse cursor instead, and undid the mouse cursor before drawing anything else on the screen and then redid it afterward. (But there were no GUI buttons; the mouse was used only to select which card to move, and other operations were done by keyboard)
22:12:12 <Zarutian> wob_jonas: so if you swinged the mouse about it actually executed a bit slower, but you could pause the fractal drawing etc.
22:13:32 <Zarutian> I am a bit sad that this thing sucumbed to a bad harddrive sector.
22:15:41 <Zarutian> when I discovered Forth I was rather pissed because Beginners' All Situations Interchange Computerlanguage was rather crappy and used the confusing mathematical notation I have come to loath.
22:16:20 <Zarutian> And here was this thing all along that was much better so long you didnt mind RPN.
22:16:44 * Zarutian hears Marvin sigh electronically
22:16:45 <zzo38> As an example of what I was asking before about OpenGL is to consider a fragment program such as this http://sprunge.us/XSHI and may want to change the name table during every frame.
22:17:22 <zzo38> (Note this program is not quite a standard ARB fragment program; it uses a preprocessor to convert the fractions with slashes into decimals.)
22:18:03 -!- Phantom_Hoover has quit (Quit: Leaving).
22:18:20 -!- mroman_ has quit (Ping timeout: 260 seconds).
22:18:25 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
22:19:02 -!- hppavilion[1] has joined.
22:21:46 -!- AnotherTest has quit (Quit: ZNC - http://znc.in).
22:22:28 <zzo38> And the program that writes it may be JavaScript using mmap.js (a Node.js package), so that is why I wanted to know the best way to upload the texture every frame. Also for sprites it may wanted to have a list of shapes to upload too every frame. (And the shader program to use would not necessarily be the one I posted; the JavaScript code may upload a different fragment program.)
22:30:26 -!- krok_ has quit (Remote host closed the connection).
22:34:33 <zzo38> Therefore, can you answer my question?
22:40:32 -!- copumpkin has joined.
22:42:23 -!- hppavilion[1] has quit (Ping timeout: 264 seconds).
22:44:24 -!- nooga has quit (Ping timeout: 276 seconds).
22:49:28 -!- nooga has joined.
23:00:59 <zzo38> Do you know that answer, please?
23:02:28 <Zarutian> int-e: said something about Buffer Objects so you could start there.
23:04:16 <shachaf> copumpkin: Do you think comments make sense on diffs directly like that, or in a code review tool?
23:04:29 <zzo38> I did read about Buffer Objects, but am unsure the best way to use them (without causing too much desynchronization and/or other problems).
23:04:31 <shachaf> I do wish more people used good code review tools.
23:04:40 <copumpkin> I want to "tell a story" in my code review
23:04:49 <copumpkin> but there's no reason not to start at diffs
23:05:47 <Zarutian> zzo38: well you have deplete me of my GPU knowledge. Perhaps ask in #OpenGL ?
23:06:25 <zzo38> Zarutian: I tried; they won't answer me.
23:09:19 * Zarutian really wished that all terminal emulators supported Sixels (They arent that complex)
23:09:36 <zzo38> Does any one other than xterm do?
23:10:27 <zzo38> Display pictures on a DEC VT terminal
23:10:39 <Zarutian> zzo38: well any other terminals that support VT340
23:11:19 <copumpkin> I'm stuck at ORD for another hour and a bit
23:11:36 <Zarutian> copumpkin: what is ORD in this context?
23:13:47 <wob_jonas> copumpkin: on the way away, or on the way home?
23:14:48 <Zarutian> copumpkin: oh, I didnt catch that from context. Like KEF or BIRK, I gather.
23:15:18 <copumpkin> wob_jonas: way home :) second layover
23:17:44 <Zarutian> airports, it reminds me. A guy I heard of took a frozen water bottle through security because he successfully argued it was not a liquid. Frankly I think the security guys just saw that was water ice and didnt bother further than putting it through the x ray machine.
23:20:52 -!- DHeadshot has joined.
23:22:15 <Zarutian> wob_jonas: btw ftp://ftp.cs.utk.edu/pub/shuford/terminal/all_about_sixels.txt should illuminate sixels
23:22:26 <wob_jonas> Zarutian: hmm, how does that work? does he pour the melted part of the water on the floor right before he gets to the front of the queue of security?
23:23:30 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
23:24:00 <Zarutian> wob_jonas: nope, wetish paper towel evaporation cooling. Though he could just have as easily just drunk the melted part before getting to the front of the queue.
23:25:08 <wob_jonas> I know the security people do sometimes let you through even with forbidden objects, I've seen and heard of examples
23:25:15 <Zarutian> the 'no liquids' crap is generally considered rather idiotic by all.
23:25:46 <wob_jonas> Zarutian: no it's not! it's actually one of the more brilliant parts of the rule
23:26:08 <wob_jonas> it makes many people buy drinks in the part of the airplane after security
23:26:37 <Zarutian> or just fill up at the drinking fountain after security.
23:26:42 <wob_jonas> just like how the hand baggage size is also checked only at security, and anything you buy after, including large wine bottles, doesn't count in it
23:26:57 <wob_jonas> Zarutian: sure, some of them do that. I sometimes do too.
23:27:15 <Zarutian> I swear those hand baggage size shrinks every year
23:28:07 <Zarutian> just heard stories from people that do
23:28:12 <wob_jonas> Zarutian: no, it's more like the atoms are getting bigger in every object because of the Hubble constant
23:29:13 <Zarutian> wob_jonas: that is not even internally consistant. Because if they did get bigger then the relative sizes of objects wouldnt change.
23:30:58 <Zarutian> it is just that airlines seem to follow the same semi-parabolic pattern, on large scale, that their airplanes do on small scale.
23:31:39 <wob_jonas> what? airplanes are shrinking too?
23:32:36 -!- DHeadshot has quit (Ping timeout: 240 seconds).
23:33:18 <wob_jonas> or... planes are free falling to the ground in a parabolic path?
23:33:58 <Zarutian> wob_jonas: I said semi-parabolic. They take off get up to crusing altitude then dive and land.
23:36:15 <Zarutian> the Hubble constant is basically saying that more space pops into existance over time. If we think of space as made up of biased smallscale randomly connected graph of planc sized cells then there is new cells coming into existance or at least connecting themselfs into the graph all over the place then...
23:37:51 <Zarutian> ... then it should be rather obvious that atoms which are mostly empty get bigger too at nearly the same rate.
23:39:03 <Zarutian> (the biasing of the graph is basically how space folds and gravity emerges)
23:41:44 <zzo38> Someone had before ask me to make up a Magic: the Gathering card "Demons In Your Nose {UB}", and now maybe they should be added into esolang wiki in article about undefined behaviour or maybe it should go only in the talk page instead.
23:42:47 <HackEgo> [wiki] [[Talk:Undefined behavior]] https://esolangs.org/w/index.php?diff=50850&oldid=35567 * Zzo38 * (+251)
23:43:02 <Zarutian> somebody used Cs undefined behaviour to lure in nasal demons and capture them?
23:45:34 -!- copumpkin has joined.
23:48:16 <shachaf> I've been to MDW but never to ORD.
23:49:37 -!- tromp has joined.
23:53:50 <zzo38> About my previous question, what I found is that MegaZeux just uses glTexSubImage2D to update the texture during each frame.
23:54:19 <copumpkin> I've never actually been to chicago
23:54:52 -!- wob_jonas has quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client).