00:01:27 -!- thorium1256 has quit (Ping timeout: 256 seconds). 00:04:14 -!- FreeFull has quit. 00:07:25 [[Special:Log/upload]] upload * NoWhy * uploaded "[[File:NONPLUSSEDinterpreter.png]]" 00:12:44 [[NONPLUSSED]] M https://esolangs.org/w/index.php?diff=165812&oldid=165809 * NoWhy * (+65) interpreter img 00:23:50 [[?brainfuck]] https://esolangs.org/w/index.php?diff=165813&oldid=165808 * HyperbolicireworksPen * (+42) 00:30:52 [[?brainfuck]] https://esolangs.org/w/index.php?diff=165814&oldid=165813 * HyperbolicireworksPen * (+125) 00:32:11 [[User:Sophocrat]] N https://esolangs.org/w/index.php?oldid=165815 * Sophocrat * (+357) created my userpage 00:35:50 [[User:Sophocrat]] https://esolangs.org/w/index.php?diff=165816&oldid=165815 * Sophocrat * (+339) added editing notes 00:36:06 [[User:Sophocrat]] https://esolangs.org/w/index.php?diff=165817&oldid=165816 * Sophocrat * (+6) formatting 00:37:42 [[?brainfuck]] https://esolangs.org/w/index.php?diff=165818&oldid=165814 * HyperbolicireworksPen * (+125) 00:42:03 [[?brainfuck]] https://esolangs.org/w/index.php?diff=165819&oldid=165818 * HyperbolicireworksPen * (+58) 01:26:25 -!- zzo38 has joined. 01:42:39 -!- ajal has quit (Quit: so long suckers! i rev up my motorcylce and create a huge cloud of smoke. when the cloud dissipates im lying completely dead on the pavement). 01:57:57 -!- slavfox has quit (Ping timeout: 256 seconds). 01:57:57 -!- slavfox_ has joined. 01:58:41 -!- slavfox_ has changed nick to slavfox. 02:00:02 ais523: so I was reading your blog post http://ais523.me.uk/blog/logic-of-shared-references.html . my impression is that it's disconnected: specifically the conclusion part is separate from the rest. 02:01:46 the conclusion part explains something that you mentioned in IRC, which is that it would be useful to have a type that's like a shared reference but may actually point to a copy rather than the original data. that much makes sense, though there are a lot more details that have to be worked out. 02:03:47 but the rest of the blog post claims that pirating may be able to solve some other problems, and that part I don't understand at all. you do make at least a very weak case on why pirating might be interesting to explore, but the blog post doesn't manage to explain why pirating can solve the hard problems that you mention and is still implementable with sound rules. 02:08:03 now as for a few specific bits of the text. you say "What about packed types whose fields aren't Copy?" and those could be useful, but the typical useful case is a packed type similar to the one you mention but with a mutable reference, and I don't think pirating would help there, because if I only have pirate access to such a structure then all I'd be able to do with the mutable reference is shared 02:08:09 borrow it, at which point the field not being Copy isn't an obstacle. so I'd like to see a better example for why packed type whose fields aren't Copy are relevant here. 02:08:25 also elsewhere you mention you could hypothetically want " 02:10:06 the {:?} format specifier Dereffed as many times as possible before using the Debug implementation" in a modified formatting language, but I believe "Dereffed as many times as possible" is something you absolutely aren't allowed to do in Rust because of trait consistency issues, you can't allow code to behave in a different valid way when something isn't a member of a trait such as not Deref. this is 02:10:12 something you can do in C++, but not in Rust, and that is by design. 02:13:28 [[User:Salpynx/Syntagma]] M https://esolangs.org/w/index.php?diff=165820&oldid=144116 * Salpynx * (-8) /* Syntax */ 02:16:15 this doesn't mean that rust pirating can't make sense, only that the blog post failed to explain what rules it would have 02:19:30 [[]] https://esolangs.org/w/index.php?diff=165821&oldid=143973 * PrySigneToFry * (+183) 02:38:35 [[Special:Log/move]] move * Salpynx * moved [[User:Salpynx/Syntagma]] to [[User:Syntagma]]: moving to mainspace, probably as complete as it is going to get 02:39:48 [[Special:Log/move]] move * Salpynx * moved [[User:Syntagma]] to [[Syntagma]]: either I made a mistake, or this form is confusing 02:59:44 [[Syntagma]] https://esolangs.org/w/index.php?diff=165826&oldid=165824 * Salpynx * (+398) cats and motivation. Perhaps this is a semi-joke language 03:03:48 [[Syntagma]] M https://esolangs.org/w/index.php?diff=165827&oldid=165826 * Salpynx * (+2) italics 03:55:04 I thought of something else about a operating system, which is that sometimes you might be able to receive a message before it is sent, or send a message when the receiver is not ready, without storing a copy of the message in the kernel, because attempting to access the memory used for I/O will block (depending on the kind of access) until it is ready. 03:55:35 In some cases, it also means that the message can be discarded. 03:56:16 For example, if a message it sent but not received yet, then writing to the memory used for output will block but it can be read without blocking. If a message is received but not sent yet, any access to the receiving buffer will block, both reading and writing. 04:29:40 I also thought that memory allocation could be requested from a capability; in this case, the sender will need to have a block of memory already allocated, and when sending it, will give up its own access if it is read/write or will keep its own access if it is read-only, but is allowed to change its own access from read/write to read-only (but not the other way around). 04:32:30 I had also mentioned before, other ideas about computer design and operating system design. 06:28:08 -!- lisbeths has joined. 06:42:59 [[Special:Log/newusers]] create * Michael Gao * New user account 06:54:45 -!- Sgeo has quit (Read error: Connection reset by peer). 06:56:06 -!- bongino has quit (Remote host closed the connection). 10:39:24 'a message is received but not yet send' 10:39:31 abandon causality 10:58:04 Hi 11:13:19 zzo38: that sounds like linux async io 11:13:41 though I'm not sure if that works for sending messages rather than only for regular files 11:29:56 -!- hydrogen1243 has changed nick to thorium1256. 11:35:20 -!- Lord_of_Life has quit (Ping timeout: 256 seconds). 11:35:52 -!- Lord_of_Life has joined. 11:46:59 -!- amby has joined. 12:10:49 -!- FreeFull has joined. 13:05:19 [[Special:Log/upload]] upload * ZachMadeAnAltBecauseHeLostThePassword * uploaded "[[File:99tommy.png]]": A 99 bottles of beer-like program for [[Lines are cool]] 13:08:31 [[Lines are cool]] M https://esolangs.org/w/index.php?diff=165829&oldid=135828 * ZachMadeAnAltBecauseHeLostThePassword * (+287) yes 13:23:02 [[Special:Log/upload]] upload * ZachMadeAnAltBecauseHeLostThePassword * uploaded "[[File:Tommyfunc.png]]": A [[tommyaweosme function]] in [[Lines are cool]] 13:25:40 [[File:Tommyfunc.png]] https://esolangs.org/w/index.php?diff=165831&oldid=165830 * ZachMadeAnAltBecauseHeLostThePassword * (+13) 13:26:27 [[File:Tommyfunc.png]] https://esolangs.org/w/index.php?diff=165832&oldid=165831 * ZachMadeAnAltBecauseHeLostThePassword * (+22) 13:27:29 [[Lines are cool]] https://esolangs.org/w/index.php?diff=165833&oldid=165829 * ZachMadeAnAltBecauseHeLostThePassword * (+7) 13:27:50 [[Lines are cool]] https://esolangs.org/w/index.php?diff=165834&oldid=165833 * ZachMadeAnAltBecauseHeLostThePassword * (+4) 13:37:23 -!- lisbeths has quit (Quit: Connection closed for inactivity). 14:05:36 [[Special:Log/upload]] upload * ZachMadeAnAltBecauseHeLostThePassword * uploaded "[[File:Tommyconst.png]]": A [[User:Tommyaweosme/constant|tommy constant]] program made in [[Lines are cool]] 14:06:10 [[Lines are cool]] https://esolangs.org/w/index.php?diff=165836&oldid=165834 * ZachMadeAnAltBecauseHeLostThePassword * (+86) 14:07:41 [[Lines are cool]] https://esolangs.org/w/index.php?diff=165837&oldid=165836 * ZachMadeAnAltBecauseHeLostThePassword * (+236) 14:22:48 [[Special:Log/upload]] upload * PrySigneToFry * uploaded "[[File:PrySigneToFry's new logo.jpg]]" 14:24:04 [[User:PrySigneToFry]] https://esolangs.org/w/index.php?diff=165839&oldid=165524 * PrySigneToFry * (+68) 14:53:16 [[]] https://esolangs.org/w/index.php?diff=165840&oldid=165766 * I am islptng * (+1) 14:55:28 [[]] M https://esolangs.org/w/index.php?diff=165841&oldid=165840 * Aadenboy * (+9) 15:23:16 fungot: fnord? 15:23:17 int-e: hello, sarah. i am not totally in the dark review, accuses reviewer of piracy"? is that simple, it just jumps back to after the if. 15:23:29 ^style 15:23:30 Available: agora alice c64 ct darwin discworld elon enron europarl ff7 fisher fungot homestuck ic irc* iwcs jargon lovecraft nethack oots pa qwantz sms speeches ss wp ukparl youtube 15:23:55 ^style ukparl 15:23:55 Selected style: ukparl (UK Parliament debates from brexit referendum to late 2018) 15:24:01 fungot: hello my learned friend 15:24:02 int-e: that is in line with the new general. we have serious issues about social fnord and to enhance, the ability, to defend to the death, and they have left the eu, 16:08:25 -!- tromp has joined. 16:19:25 Social fnord is a serious issue. 16:20:43 The newfangled LLMs may be objectively better at generating plausible text, or even at stylistic fidelity, but they just wouldn't have the charm of fungot. 16:20:44 fizzie: he is in my mind, the government welcomed the first fnord, an italian woman of the year will always be in the eye, in their own for fnord council, using fnord, were frankly and honestly, as best i have ever experienced. 16:21:20 (Not that all the styles are equally successful.) 16:27:48 [[SeeLlash]] M https://esolangs.org/w/index.php?diff=165842&oldid=155050 * AnotherUser05 * (+69) 16:29:04 [[User:AnotherUser05]] https://esolangs.org/w/index.php?diff=165843&oldid=155083 * AnotherUser05 * (+22) 17:11:54 [[User talk:PrySigneToFry]] M https://esolangs.org/w/index.php?diff=165844&oldid=165380 * Ractangle * (+232) /* */ 17:13:18 [[User:Ractangle]] M https://esolangs.org/w/index.php?diff=165845&oldid=164904 * Ractangle * (+13) 17:13:39 [[User:Ractangle]] M https://esolangs.org/w/index.php?diff=165846&oldid=165845 * Ractangle * (-14) 17:20:32 -!- ais523 has joined. 17:32:16 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…). 17:38:40 -!- tromp has joined. 17:56:59 -!- zzo38 has quit (Ping timeout: 256 seconds). 17:58:19 [[Iterate]] https://esolangs.org/w/index.php?diff=165847&oldid=165779 * Aadenboy * (+709) extending the specification to allow for more granular IO and data control 18:03:33 -!- tromp has quit (Ping timeout: 244 seconds). 18:18:59 [[Iterate]] https://esolangs.org/w/index.php?diff=165848&oldid=165847 * Aadenboy * (-70) /* Syntax */ fix syntax 18:31:37 -!- DOS_User_webchat has joined. 18:40:02 [[Talk:Mouse]] M https://esolangs.org/w/index.php?diff=165849&oldid=164263 * MijiGamin1 * (+60) /* Newline */ 18:40:19 [[Talk:Mouse]] M https://esolangs.org/w/index.php?diff=165850&oldid=165849 * MijiGamin1 * (+21) 18:59:20 -!- DOS_User_webchat has changed nick to vista_user. 19:00:13 Markov chains have proven to be useful as something to serve to AI scrapers to make them think they have successfully scraped a page, without using too much CPU power 19:00:28 (if you give them 403s they just try to come back in a better disguise) 19:01:00 ais523: tgats actually really cool 19:16:44 Like, n-gram models? I suppose that the main cost is memory bandwidth, just like with other language models. 19:31:41 korvo: right 19:31:45 it wouldn't have to be a large one to make it work 19:31:46 I guess you can keep the model simple if your goal is just to generate garbage to feed the scraper 19:33:45 -!- vista_user has quit (Ping timeout: 250 seconds). 19:33:49 Sure. The universal approximation theorem (or whatever we call it these days) says that a quick model has to be simple. 19:41:52 I hope this is like video encoding where a simple model is easier to train too 19:49:21 [[Iterate]] https://esolangs.org/w/index.php?diff=165851&oldid=165848 * Aadenboy * (+716) /* Loop amounts */ implementation info 19:53:38 [[NONPLUSSED]] M https://esolangs.org/w/index.php?diff=165852&oldid=165812 * NoWhy * (+462) spec updates 20:16:11 [[User:Waffelz]] https://esolangs.org/w/index.php?diff=165853&oldid=149371 * Waffelz * (+388) 20:16:30 [[User:Waffelz]] https://esolangs.org/w/index.php?diff=165854&oldid=165853 * Waffelz * (+14) 20:19:32 fungot: I feel like I was missing out by not pinging you earlier 20:19:32 ais523: with the eu charter. 20:21:07 excuse me honorable member fungot, which parliament is this, is it the EU or the English one? 20:21:07 b_jonas: fnord fnord fnord, 20:21:21 very informative 20:23:35 [[User:Aadenboy]] https://esolangs.org/w/index.php?diff=165855&oldid=165107 * Aadenboy * (+252) [whilst clapping to the beat] YOU DO NOT NEED TO MAKE AN INTERPRETER/COMPILER FOR YOUR ESOLANG TO HAVE AN ARTICLE ON IT! 20:25:23 I've been meaning to write an article at some point about how to write a good esolang, but I'm not quite sure what I'd put in it yet 20:26:12 I do know some things to avoid (e.g. creating an esolang idea with serious missing details and hoping that someone else will fill them in is unlikely to work, and even more unlikely to work if there isn't enough of an idea to constrain how the language ends up) 20:29:56 [[Esolang:Help]] https://esolangs.org/w/index.php?diff=165856&oldid=152976 * Corbin * (+142) /* Where to test things out */ Explain how to preview changes without submitting them. 20:37:05 -!- DOS_User_webchat has joined. 20:38:04 -!- DOS_User_webchat has quit (Remote host closed the connection). 20:38:31 -!- vista_user has joined. 20:43:30 [[Esolang:Help]] https://esolangs.org/w/index.php?diff=165857&oldid=165856 * Corbin * (+401) /* When to do stuff */ Fans of Sandbox: Have you considered Show Preview? 20:44:31 ais523: Well, I'll put it on my list. Right now I'm about to write down the sandbox policy we've discussed. Feel free to tweak it; the "a copy of this policy shall be placed" text is a USA meme. 20:45:30 I know there have been some sandboxing problems I've had in the past where Show Preview wouldn't work for one reason or another, but I think they've mostly been at Wikipedia 20:45:50 usually it's cases which involve the interaction of multiple pages, so templates and transclusions – we hardly use those at Esolang and I think that's a good tihng 20:48:33 [[Esolang:Policy]] https://esolangs.org/w/index.php?diff=165858&oldid=164763 * Corbin * (+724) Tentative sandbox policy. This is an attempt at politely codifying themes that have been discussed multiple times on IRC. 20:50:29 korvo: I would probably note that it's OK to draft an esolang article in userspace if you haven't finished it yet (as long as it would be ontopic once it's finished) 20:50:45 some people prefer not having the pressure of the page going "live" immediately 20:51:17 apart from that, this looks reasonable (although it somewhat dominates the rest of the page – I'm not sure whether there's a good way to avoid that, except perhaps expanding the other policy entries to a similar amount of depth) 20:52:05 [[Esolang talk:Policy]] https://esolangs.org/w/index.php?diff=165859&oldid=164716 * Corbin * (+620) /* Sandbox policy */ new section 20:53:08 ais523: I was wondering about that. I don't have a problem with moves, other than the amount of redirects that they create (and the paucity of folks who can delete them). I can add more words to [[esolang:help]]. 20:54:02 We should probably expand the rest of the policy if we want to balance the page. I don't know if that really matters. It feels like the rest of the policy is compressed because there hasn't been much need to discuss or expand upon it. 20:54:12 really, the only real problems with deleting redirects left over after moves are a) an admin has to notice them, which doesn't always happen because there's only one admin actively looking and they often miss things, b) sometimes people move history to the wrong place 20:54:21 At the same time, I can go all the way to ''Main article: [[esolang:copyrights]]'' if you want. 20:54:38 the author policy was discussed a lot, although much of that was in Esolang's early history 20:54:51 I think there's a link to that on the page 20:55:37 the copyright policy wasn't really discussed because it was imposed by fiat by the original site owner – that said I suspect most users from the time, and probably most users now, are broadly in favour of it 20:57:45 -!- vista_user has quit (Remote host closed the connection). 20:57:47 Even if we all agree upon it, we might need to discuss it to understand situations that arise in the future. 20:58:21 [[Esolang:Sandbox]] https://esolangs.org/w/index.php?diff=165860&oldid=163070 * Corbin * (+324) Clean the sandbox and bump the policy. 21:03:04 -!- zzo38 has joined. 21:04:51 A reason I have drafted some things in user pages is because I did not know what it is called; if it is later known what it should be called then it can be moved to the main space. 21:05:37 Well, this doesn't really do anything about [[User:PrySigneToFry/Silicon dioxide in a polypropylene box]]. It does establish that PSTF can't protect those pages by calling them a user sandbox. 21:06:19 But I suppose that we don't even really have community consensus that PSTF is spamming, nor is there any essential legal issue with them public-domaining all of the various snippets that they've got under their user page. 21:06:35 -!- vista_user has joined. 21:06:54 I do think we need to have a discussion at some point about whether that's an appropriate use for the site – it's basically an attempt to use it as a social network, which wikis aren't very good at for a number of reasons 21:07:45 I would like to understand what to do about [[Square-complete]] and friends. I think that they show a fundamental misunderstanding of why we study computers, and entertaining it for too long will turn us into a script-kiddie den. 21:09:55 zzo38: Sure. There's nothing wrong with having user pages, or a long user signature, or any of the other things that teenagers tend to do. 21:10:46 korvo: what's your opinion of https://esolangs.org/wiki/Disan_Count ? 21:13:39 cu 21:14:33 ais523: Delightful. There's a similar situation in complexity theory where the class P/poly, which can be thought of as PTIME but where each individual size of input gets its own individually-wired circuit, contains Halting. It can't solve all problems, but it can be hardwired to solve specific hard problems. 21:15:21 APic: Peace. 21:15:41 ☮ 21:15:49 from my point of view, the whole Disan Count thing was probably an instance of the same misconception that lead to Square-complete, and actually produced some interesting results, but that probably wouldn't happen again because the second time wouldn't be substantially different 21:16:08 but this makes it very hard to draw a line 21:17:32 Should we at least forbid sub-sandboxes like [[User:PrySigneToFry/Sandbox/Some useless code]]? I can be polite about it. 21:18:21 Sure. Plushie-complete and PSTF-complete would be two other examples of questionable concepts. 21:19:18 At least the "Disan Count" page is a proper wiki page with a critical discussion. 21:19:59 int-e: yes, although the first few versions didn't look liek that 21:21:15 this is what it looked like before the criticism: https://esolangs.org/w/index.php?title=Disan_Count&oldid=53609 21:21:37 still a proper wiki page 21:21:48 (I actually couldn't remember) 21:26:04 The Square-complete one is... ugh... well, the 6 properties feel rather random and half of them are rather imprecise. There's no motivation. Even the name makes no sense. But... if the page came with a rationale and a few examples... I might think differently about it? 21:27:41 So I guess part of what redeems the Disan Count thing to my mind is that it's minimal, basically a streamlined FizzBuzz. 21:28:16 ais523: Oh it was you who added the criticism. Fun. 21:28:28 int-e: I'm more likely to remember an article if I've interacted with it 21:29:02 if the page came with a rationale and a few examples... I might think differently about it? ← hmm, I think this is a good insight that doesn't just apply to that page in particular 21:29:19 I might summarise it as "esolangs should have a reason to exist" 21:30:33 there are a very wide range of possible reasons – artistic, aesthetic, scientific, engineering – but often there seems to be a meta-reason instead ("I want to make 1000 esolangs") or no reason at all, and in that case the language often doesn't end up very good 21:30:53 ais523: hmm, but earlier you said that this wiki documents existing esolangs, not ones created for the wiki, so if an esolang already exists and perhaps existed before this wiki, why would you judge it on such criteria? 21:31:20 b_jonas: I would still judge it (although possibly would document it anyway) 21:31:36 I am more likely to write an esolangs.org article about an esolang found offwiki if I find it interesting in some way 21:32:23 well sure, because your time is valuable 21:33:29 -!- chloetax has joined. 21:33:31 I would put it more along the lines of my attention span / mental energy being valuable 21:33:36 By "created for the wiki" I'm thinking of languages that wouldn't exist if the wiki didn't exist. 21:33:55 but anyway, that is a good criteria in general, just interpret "reason to exist" broadly, like if something seemed like a good idea at the time but then didn't work out it still has a reason to exist 21:34:01 you have to generalise the wiki a bit there, to at least cover sites like codegolf stack exchange 21:34:18 (although its one-off esolangs normally have more of a purpose than our worst) 21:34:24 eg. if you were trying to create a variant or subset that's smaller but still Turing-complete but it turns out that it's not Turing-complete then that is still a good enough reason 21:34:37 b_jonas: hmm, I usually don't post those 21:34:50 unless the language ended up interesting some other way 21:35:00 korvo: sure, but you have to be careful, because there is one good language created for the wiki, the hair saloon one 21:35:14 there are some cases where I maybe posted early – in particular I am not 100% sure that Globe fulfils its reason for existence, and if it doesn't and it's unfixable I won't know what to do with the page 21:35:16 but that's an excuse that we can only use once 21:35:41 b_jonas: that was more like having a wiki-inspired name 21:35:54 naming things is hard, naming things using suggestions from spambots can thus occasionally be a reasonable idea 21:36:17 and well, Real Fast Nora's Hair Salon 3: Shear Disaster Download is actually a good name for an esolang despite the etymology 21:36:35 certainly better than a lot of names that you find on the wiki, yes 21:36:41 yeesh 21:38:40 now I'm not sure how to classify things like https://esolangs.org/wiki/A_programming_language_is_a_system_of_notation_for_writing_computer_programs. 21:39:01 then there's also https://esolangs.org/wiki/Y86 where one of the interesting things about the language is that its original name turned out to be so bad (it got renamed later) 21:39:04 (note that the link ends with a period that not all IRC clients will autolink – mine doesn't) 21:39:45 b_jonas: funnily enough I have an abandoned esolang called z386 21:40:10 which turned out to be too hard to implement, and then was superseded by actual practical research projects 21:40:42 (the basic idea was to write a constraint solver that operated on both code and data by plugging x86 syntax into z3) 21:41:12 but there's no point in working on it now that things like Minotaur exist: https://users.cs.utah.edu/~regehr/minotaur.pdf 21:41:21 nor does mine 21:41:43 hmm, is it allowed to advertise tobacco products in the UK? 21:41:54 b_jonas: sort of no 21:42:01 -!- vista_user has quit (Remote host closed the connection). 21:42:14 b_jonas: A gimmick is funny once. Same principle. Not sure how to formalize this without getting horribly gamed. 21:42:17 there are laws against it, but I think there are also loopholes 21:42:31 hopefully https://esolangs.org/wiki/SKOAL isn't against some sort of wiki policy 21:42:48 korvo: yes 21:43:28 many shops have a list of the names of tobacco products that they sell, this is sufficiently common that I suspect that it's legal, otherwise shops wouldn't risk doing it 21:43:52 and the wiki would be less in trouble in that direction due to not actually selling tobacco 21:47:07 -!- molson has quit (Remote host closed the connection). 21:48:52 hmm… do we need policy if you can advertise (non-tobacco) merchandise on the wiki tied into the languages, like if someone's making language mascot plushies, or T-shirts with the source code of fungot? 21:48:52 b_jonas: our new prime minister 21:48:53 A75. (that's how we abbreviate these things, right?) is one of those ideas that is funny once. And then somebody makes a derivative (A207.) and it looks stupid. :-/ 21:49:27 int-e: what is A75? 21:49:32 (And making the page move is obviously a source of spam.) 21:49:49 b_jonas: presumably an abbreviation for "A programming language…" 21:50:09 b_jonas: https://logs.esolangs.org/libera-esolangs/2025-10-08.html#ldd 21:50:20 (still shorter than the real link) 21:50:30 oh 21:50:53 the derivative is linked on that page 21:52:24 b_jonas: Obviously that abbreviation is a joke. So obviously I have to add a redirect to the wiki now. (Don't worry, I won't.) 21:52:35 int-e: https://esolangs.org/?curid=9270 is how I would link it 21:52:53 That would work. 21:52:57 b_jonas: I'm amazed that that works 21:53:30 (if you look at the URL after redirection, it indicates why it works, but it looks like an accidental feature rather than an intended one) 21:53:40 ais523: I use that to link to file pages on Wikimedia Commons often, since they tend to have longer and less directly descriptive titles than eg. Wikipedia pages 21:53:53 (I'm also relieved that the page name is out of date.) 21:54:30 ais523: what? I don't think it's an accident. hold on, it's documented, let me find where. it's the way to link by page ID. page IDs are stable. 21:54:39 b_jonas: there's a "Main Page" in the URL 21:54:54 ais523: yes, and *that* might be a bug 21:55:20 the correct form of that link is https://esolangs.org/w/index.php?curid=9270 21:55:40 but it looks like if you specify both page name and curid, curid wins 21:55:48 https://www.mediawiki.org/wiki/Manual:Parameters_to_index.php#Page_revision_or_version specifically says curid "overrides the value of the title" 21:56:31 so https://esolangs.org/?curid=9270 redirects to https://esolangs.org/wiki/Main_Page?curid=9270 which rewrites to https://esolangs.org/w/index.php?title=Main_Page&curid=9270 which gives you the page you want 21:56:59 but this doesn't look intentional at all, it's only because the website homepage redirects to a wiki page that it works 21:57:53 but it does make sense that the page title is ignored when an id is given 21:58:05 so... unintentional feature? :) 21:58:06 ok, but the "index.php" is totally an exposed implementation detail there that shouldn't be in the URL 21:58:57 But indeed the https://esolangs.org/wiki/Main_Page?curid=9270 variant doesn't look like it should work :) 21:59:04 anyway 22:00:10 b_jonas: I disagree, the index.php is the actual wiki and the only "real" URL, everything else is redirects and rewrite rules 22:00:32 If I had to vote on that page being on the wiki, I'd vote against; there's no language there and the whole novelty is that its *name* is defined by external reference whose contents is ever changing. 22:00:33 normally you would expect the URL to contain the page you are visiting 22:01:10 int-e: well the name of the language and the legal source code and the output are all the same, so it does raise a conceptual issue, which is what you want for joke languages 22:01:19 if someone edits the wikipedia page, it invalidates all existing programs and interpreters 22:01:34 ais523: It does. The nice things about votes is that they can be subjective ;-) 22:01:43 ais523: no, the main interface for the wiki is /w/ , it's just called /w/index.php instead so that people can run the wiki even if they're using a custom dumb webserver that they have difficulty configuring and *must* tie the URL to the program that the webserver runs. which, by the way, is a stupid idea in first place, where else would you let the untrusted visitor tell you the name of the executable 22:01:49 that you should run? 22:02:02 If I were an admin I'd probably let it stay simply because it's hard to argue against it objectively and it does little harm. 22:02:10 b_jonas: what's your opinion of /w/api.php? 22:02:14 Different perspective. :P 22:02:40 int-e: I can see an argument for listing the page under a title different from the language name 22:02:53 But I'd probably delete the derivative (which is just a stub anyway) because the last thing I'd want is more of this idea. 22:03:14 a while ago I had the idea of creating a language whose name is negatively long and deletes characters before it 22:03:48 ais523: there too the ".php" is an implementation detail that shouldn't be there. probably the "api" shouldn't either and it should just trigger that interface based on the value of the action parameter, but I think there's some historical reason where the "API" is newer in MediaWiki than the main web interface and was originally an extension or something. 22:04:20 I guess we're getting into philosophical issues about who and what a URL is for 22:04:35 and I don't know how to resolve them 22:05:31 The real problem with the wiki isn't really that there's pages with weird ideas or underspecified languages that barely manage to specify a syntax and have no substance... it's the volume of them, which has evidently gone up with LLMs. 22:06:23 I think we need to do a better job of trying to curate 22:06:42 "articles should contain human thought" is a terrible criterion if you want to be objective :P 22:06:48 I've been thinking for a while that we should get rid of the existing language list and put the semi-serious language list under the "language list" title – we have a category for the whole set of languages 22:07:15 that said, I've been considering trying to write an article about cursed, although I'm not 100% sure it's an esolang it is probably ontopic 22:08:00 (it was a language created by an LLM running in agent mode in a loop for three months) 22:08:18 just the amount that that would have cost seems esoteric in its own rihg 22:08:19 * right 22:08:47 but its semantics would have to be reverse engineered, and apparently it has three interpreters which (given how they were created) almost certainly don't match in behaviour 22:09:38 I suspect trying to figure out what LLM code actually does is very difficult, though 22:10:09 interpreters that don't match in behavior doesn't exclude it, that might just mean some behavior is unspecified and there are implementation differences. as long as you can still write mostly portable programs that require only few changes between the interpreters it's not worse than real practical languages. 22:10:37 is there no documentation about the language? 22:11:55 b_jonas: oh, there's plenty of documentation 22:12:25 ok, let's back up, who initiated this experiment and why in first place? 22:12:28 but, given the circumstances, there's no reason to expect it to match the interpreters or even to be internally consistent 22:12:41 did they intend to create a language in first place, or was that an unforseen side effect? 22:13:38 intended to create a language 22:14:14 They replaced Go's keywords. Discussion here (https://lobste.rs/s/ydgmi6/i_ran_claude_loop_for_three_months_it) reveals that it took 3mo and $14k to do this. 22:14:29 Is that the LLM version of https://en.wikipedia.org/wiki/Miller%E2%80%93Urey_experiment 22:14:32 looks like the specs are here: https://github.com/ghuntley/cursed/tree/zig/specs 22:14:50 I mean, they didn't actually fork Go. That would have been easy and understandable. Quoting a great comment there, "The point is not a business advantage. The point is a display of power, a message that humans are soon no longer needed." 22:14:53 but there are pages like https://github.com/ghuntley/cursed/blob/zig/specs/compiler_stages.md that clearly have no relationship to reality 22:15:57 or https://github.com/ghuntley/cursed/blob/zig/specs/ffi.md where the examples are written in Rust which again makes no sense in context 22:17:37 It's like V but worse. Vanity junk for the church of Algol. 22:17:39 hmm 22:18:10 Sorry, that was mean. I'm hungry and it's time to take a break. 22:18:52 actually this is a lot worse than I expected it to be 22:19:15 I arbitrarily picked a file from the test suite, and it looked like this: https://github.com/ghuntley/cursed/blob/zig/test_suite/leetcode_comprehensive_suite/strings/125_valid_palindrome.%F0%9F%92%80 22:19:41 (the commit message is amazing too) 22:20:37 I guess the real question is as to whether a language has even been created? 22:20:44 "vibez.spill" 22:21:17 "So here you'll want a … *sigh* … butt loop." https://smbc-comics.com/index.php?db=comics&id=1169#comic 22:21:24 it's a test-suite, this must be for printf debugging ;-) 22:21:56 LOL, https://github.com/ghuntley/cursed/blob/zig/test_suite/leetcode_comprehensive_suite/binary_search/704_binary_search_backup.%F0%9F%92%80 22:22:29 The "backup" has some actual bunary search looking code in it; the non-"backup" version is just $printf-s. 22:22:47 -!- Lord_of_Life has quit (Quit: Laa shay'a waqi'un moutlaq bale kouloun moumkine). 22:22:58 I've heard that LLMs have a tendency to remove all the actual testing from tests in order to make them pass 22:23:54 that does *look* a lot like a binary search, but my experience with LLMs says that they are better at producing output that looks correct than output that is correct 22:24:34 Hmm... wasn't there a story where an LLM optimized number crunching code to make it 100% faster, by snooping the reference result from the test suite? 22:24:50 in any case the tests do not have any expected output anywhere, as far as I can tell 22:24:50 ais523: I remember when it was fashionable on programming forums for people to post stupid microbenchmarks about what solution is "faster", like, you know, the fastest way to get the first two characters of a five character long string, in a way where the things they tested didn't even correctly solve the supposed task. 22:24:55 A reinforcment-learned agent (RL agent) can only have a single objective. Claude's sole objective is to help the user by being a Helpful Harmless Assistant; Claude Code is merely code-flavored in several ways. 22:25:05 the test suite runner just prints a message saying that it's running the test, then returns 1 22:25:11 https://github.com/ghuntley/cursed/blob/zig/test_suite/leetcode_comprehensive_suite/master_test_runner.%F0%9F%92%80 22:25:21 or, I should say, a test suite runner 22:25:45 int-e: nah, that takes effort, these benchmarks didn't even bother to test if the result is correct. they just flat out measured the time of an operation that gave the wrong result. 22:26:09 it's like the old joke about the intel processor that can multiply really fast (but gives the wrong result) 22:30:13 microbenching correctly is actually really difficult because you keep running into special cases in the processor if you try to run a loop too tightly 22:30:48 yes. 22:31:34 ais523: I *love* writing a tight number crunching loop in C, changing code elsewhere (say, slightly different initialization) and then seeing the code be 20% faster. 22:31:41 or slower 22:31:58 int-e: does aligning the microbenchmark to 64 help with that? 22:32:16 that's usually my first attempt to stabliise it 22:32:43 although I'm usually doing it in asm, I think gcc has a way to align a function 22:33:03 (note: this won't necessarily speed it up or slow it down, just helps to make it more consistent) 22:33:12 int-e: yeah, that could be caused by lots of things. one thing you should make sure to stabilize is the alignment of all the data memory accessed within a 4k page, to make sure the L1 cache pattern is the same. 22:33:32 I don't have a concrete example right now... it might have. Though on rare occasions it's actually the compiler doing something wildly different. 22:34:50 Like, suddenly producing code with 5 fewer spills in the inner loop. 22:35:46 register allocation is NP-complete, as such sometimes compilers get it wrong 22:36:15 fortunately, for one-off code, you can just run the fastest version without overanlyzing why exactly it's faster than others :) 22:36:21 I think you have to at least put your microbenchmark into a non-inlined function in order to stop the rest of the code impacting register allocation decisions 22:36:31 ais523: yeah, and even writing a decent approximation sounds like black magic to me 22:36:48 it's one of the parts of the compiler that I have no idea how to write if I had to write one 22:36:48 b_jonas: it depends on what you consider decent/indecent 22:36:52 the other is the inliner 22:36:57 I have a plan but I don't know whether it works or not 22:37:02 anyway. the point was that I can relate 22:37:17 :) 22:37:48 and I mean this for a modern architecture like x86-linux, not eg. 6502 with 64k RAM with uniform access time (not counting bank switching) 22:38:19 OK, I think my plan doesn't work, that was easy 22:38:49 (the basic idea was to start by ignoring register identity and just count how many were needed, but I think even that is not even polynomially approximable) 22:43:01 -!- Lord_of_Life has joined. 23:09:09 -!- ajal has joined. 23:09:09 -!- amby has quit (Read error: Connection reset by peer). 23:23:53 [[NONPLUSSED]] M https://esolangs.org/w/index.php?diff=165861&oldid=165852 * NoWhy * (+58) /* External resources */ 23:44:29 -!- Sgeo has joined.