Time Nick Message 12:19 MinetestBot 02[git] 04sfan5 -> 03minetest/minetestmapper: Add cautionary note about sat_mul() 13e14f27f https://github.com/minetest/minetestmapper/commit/e14f27f41268a11c34299a8b94a380b28c6b71e9 (152023-08-08T12:18:58Z) 12:38 Abiblima hey 12:39 Abiblima worlgen is kinda stupid 12:39 Abiblima mapgen 12:39 Abiblima a minetest player places only a few hundred nodes in their lifetime as a player 12:40 Abiblima maybe at most a few thousand 12:40 Abiblima mapgen spams millions of nodes 12:40 Abiblima that will mostly never be dug or seen 12:41 Abiblima its also the no.1 performance drain mapgen 12:41 Abiblima loading all that pre-gen crap 12:43 Abiblima could just generate a surface layer mapgen 12:43 Abiblima as long as you know the seed you can pop it into voxels 12:44 Abiblima just generate an efficient land level mesh for each mapblock and pop it to voxels when dug, not placed 12:45 Abiblima either way it makes no sense to create a mapgen that spams nodes that degrade performance beyond that of a player 12:45 Abiblima who's playing the game, mapgen or the player? 12:45 Abiblima if it's the player then why is 90% of the game resources given to mapgen? 14:08 Blockhead256 Abiblima: That sounds okayish for purely heightmap 2D generation but consider caves and overhangs 14:29 MTDiscord "a player only places a few hundred nodes in their lifetime" ... "maybe at most a few thousand" ... meanwhile I'm looking at a handful of players who are in the hundreds of thousands and into the millions per server across 3 or 4 servers... 15:07 Abiblima for a player to have placed over a million nodes in their lifetime is unrealistic 15:08 Blockhead256 never underestimate what people will do with their time 15:08 Abiblima hundreds of thousands after years is possible excluding worldedit 15:08 Abiblima mathematically speaking 15:08 Abiblima over a million is unlikely 15:08 Abiblima highly highly unlikely 15:10 Blockhead256 at 1 node/s, relatively slowly, it only takes 277 hours to place that many nodes 15:10 Abiblima for example a player would need to place one node every 3.5 seconds for one and a half months 15:11 Abiblima yeah thats not going to happen 15:11 Blockhead256 people often log in excessive of 1000 hours on some games 15:11 Abiblima no one is going to place a node every second for 277 hours straight 15:11 Abiblima its highly unlikely 15:11 Blockhead256 10,000 in some MOBAs. They don't do it in one sitting, but they spend multiple hours a day 15:12 Abiblima im sure bigfoot exists too 15:12 Blockhead256 Give people a year, they'll do it easily. Many won't even consider that burned out, just a hobby. Especially with a leaderboard. Give people a leaderboard and someone will spend all their free time on it 15:12 Abiblima no one is placing over a million in a year legitimatly 15:13 Blockhead256 never underestimate the power of unemployment 15:13 Abiblima its not even methematically possible 15:13 Abiblima they'd die 15:14 Abiblima even after 10 years of playing id be surprised to see a million, but after 10 years its plausibkle 15:14 s20 ~~Let them~~ 15:14 Abiblima a million in a year just is not possible 15:14 Abiblima thats one node every 30 seconds 15:14 Abiblima even if you took breaks and placed one node every second 15:15 Abiblima you'd still not reach a million by the end of the year and be alive 15:15 s20 You mean to say Minetest mapgen generation causes...? 15:16 Abiblima im just saying mapgen eats the most resources to place more blocks than an entire server could ever dig and place 15:16 Abiblima doesn't make any sense 15:16 cosmician [Test message - changed name from s20 to cosmician] 15:16 Abiblima its like the game is playing itself 15:16 Abiblima and the players are secondary 15:17 cosmician What do you think can solve it? 15:17 Abiblima just pre-bake only visible mapblocks 15:17 Blockhead256 You'd only have to spend an hour each day 15:17 Abiblima an hour placing a node every second fullstoo? 15:17 Abiblima most players dont place a node every 5 minutes 15:17 s20 I don't know, just guessing, but wouldn't prebaking them cause worse gameplay in that moment? 15:17 Abiblima let alone every second 15:18 Blockhead256 https://paste.debian.net/1288285/ 15:18 Blockhead256 It'd be a grind to log an hour on Minetest every day but you could do it 15:18 Abiblima nah just pop them back to voxels when there is a dig action in the prebaked volume 15:18 Abiblima atm when you fly around mapgen goes nuts 15:18 Abiblima filling the db 15:19 Abiblima for what 15:19 Abiblima for nothing 15:19 Abiblima most of those will never be mined 15:20 celeron55 that idea has come and gone many times, but nobody has made a proof of concept. are you the one to do it? 15:20 Blockhead256 I'm not understanding you. It's the CPU that gets hammered once and the DB that gets bloated, or it's the CPU gets hammered every time someone travels through an unloaded area. Are you proposing a third way? A baked mesh you say? A mesh that.. needs the CPU to run normal mapgen to mak eit 15:22 celeron55 one thing is certain: a mapgen that has to come up with a surface and nothing more will be much more boring than one which can do anything. you can't e.g. generate the surface details of a cave entrance without actually generating the cave. you simply don't know how it will end up shaped like 15:23 Abiblima thats true 15:23 celeron55 cave entrances will be severely limited. and then, what do you define as the surface at the sea or lakes? 15:23 celeron55 is it the water surface? the bottom of the ocean? everything in between? 15:23 celeron55 what about floating islands? 15:24 Abiblima it gets complicated fast i see your point here 15:24 celeron55 what if you go underground and spend time there? do you lose all the optimization? is it worth it to optimize at all then? 15:25 celeron55 it will always be a compromise in some way 15:26 Blockhead256 I don't know exactly what Skamiz Kazzarch has been doing with mg_tectonic lately, other than looking for approximations and maybe some performance improvements. I wouldn't rule out that there may be approximations of our full-featured mapgens that are adequate for far-off terrain or something like that 15:26 celeron55 making a mapgen that can provide variable level of detail for quickly generating lots of land for long distance rendering would be useful 15:26 celeron55 or, well, even better, extending the current ones to do that 15:27 Blockhead256 especially with advances in training artificial neural networks, maybe one could make a very fast approximation. Or perhaps that's not much better than lower samples of Perlin noise? 15:27 Abiblima interesting lower resolutions blocks for further distances? 15:27 Blockhead256 look into the history of "farmesh"es 15:28 Abiblima https://forum.minetest.net/viewtopic.php?t=4047 15:29 celeron55 sampling noise more roughly is easy to do and would probably already satisfy the requirements of many players, including me. maybe you don't even have to do biomes at all, just render it in a foggy blue color 15:29 celeron55 however 15:30 celeron55 it can get tricky stiching that in with things built by players when they also appear at the distance 15:30 Blockhead256 Anyway this gets away from the proposal that Abiblima was actually talking about, which is saving space in the database.. 15:30 celeron55 with some luck, it won't be too jaggy 15:31 Blockhead256 some kind of interpolation from the high res blocks to the "mipmap" mesh? 15:32 Abiblima i will lookup posts and git issues/push by Abiblima 15:32 Abiblima post by ruben wardy 15:32 Abiblima https://forum.minetest.net/viewtopic.php?p=166444#p166444 15:32 celeron55 it's not exactly easy to temporarily generate parts of the world and then proceed to throw those away due to the gliches that would be generated. the way trees and stuff work is they can appear on a neighboring mapgen chunk and will be saved there. but then if you remove the generated chunk, you don't know which parts of the neighboring chunk (which may have player modifications on it) is part of that 15:33 Abiblima farmesh was removed after v5 due to incompat with v6 and bugs" - so there was farmesh at some point for long distance rendering (not space saving) 15:33 celeron55 again, you can avoid it by dumbing the mapgen down, or by adding some complex earmarking of nodes to belong to a certain mapgen chunk even while not existing in one. but again, you see, it's not so simple 15:34 Abiblima ill be back in 20-30 mins 15:34 Blockhead256 reminds me of the latest Mineclone2 release notes mentioning that the size of their artificial structures had to be smaller than in Minecraft to fit inside a mapgen chunk 15:34 celeron55 the farmesh you're talking about was completely heightmap noise based and awful looking, and didn't use any space 15:34 celeron55 it was so awful i think less than 1% of players would accept it 15:35 celeron55 v6 uses 3d noise, so it cannot produce a heightmap 15:35 celeron55 wait, or was it v5 which used 3d noise 15:36 celeron55 ah no, i think v6 uses so much brute force carving of landscape features that it cannot produce a heightmap 15:36 celeron55 i'm fairly sure v5 can't produce a heightmap either 15:37 celeron55 but that feature of v6 was tuned down at some point as some people didn't like it. it used sort of of huge partially buried caves to make surface features 15:37 celeron55 -of 15:37 Blockhead256 hmmm 15:37 Blockhead256 what mapgen will make it in the C++ next I wonder? 15:38 Blockhead256 into the master branch I mean 15:38 celeron55 anyway, none of this is very easy to do in a way that would keep players happy. it's annoying when people assume things are easy or obvious when they aren't 15:38 rubenwardy please take any posts I made 9 years ago with a huge grain of salt 15:38 Blockhead256 I would say that should go for just about anybody.. 15:39 Blockhead256 especially someone well before they became a core dev, of course 15:40 Blockhead256 Since heightmaps were mentioned, the `heightmap` mapgen object is available to every map generator, right? It's just a convenience/pre-calculation? 15:40 Blockhead256 .. and not something related to heightmap troubles with farmesh 15:41 celeron55 well, i'm just as intrigued by that as you. what do the api docs say? 15:42 celeron55 it says "Returns an array containing the y coordinates of the ground levels of nodes in the most recently generated chunk by the current mapgen." 15:44 Blockhead256 yeah so my guess is the C++ calculates it, either all the time or just when you ask for the object. It'd probably do it in a much faster way than Lua even if it's post-generation? 15:44 celeron55 looks like generally it's not noise based, it's probed node by node after the mapgen has generated its thing. but it's up to the mapgen to generate the heightmap, so if it was a simple 2d noise mapgen, it could give its base noise as the heightmap 15:46 Blockhead256 I'm sure you could make an antagonistic mapgen that makes the heightmap unsuitable. A world of 1-2 node thin slices with air in between, where the air gap varies between layers maybe for more interest. 15:47 celeron55 i kind of feel i should give a try making an as fast as possible while still kind of interesting mapgen, just to see how it would feel like 15:47 Blockhead256 then try to place a village on the 'top' of 'surface level' nodes and get some really weird placement, like the village disappears and continues on a lower level 15:47 Blockhead256 how fast is fractal? 15:48 celeron55 you don't need to make an antagonistic mapgen to make the heightmap require essentially a brute force operation like now. you just need to use 3d noise 15:48 celeron55 3d noise makes very interesting results, which is why it's used. and it's essential for caves 15:48 Blockhead256 I'm not talking about antagonising the algorithm for heightmap, but antagonising mods that depend on notions like "top of heightmap = good placement surface" or something 15:50 Blockhead256 I feel like all the caves in Minetest feel the same regardless of mapgen. Are they? 15:50 celeron55 well, it's probably still faster to produce it for mods than to have the mods essentially calculate it themselves. you can probe multiple points in the heightmap to check the ground "quality" 15:50 calcul0n_ yes, they all use the same cavegen 15:50 Blockhead256 I feel like there are some missed opportunities for things like a honeycomb or the inside of a bird bone 15:51 celeron55 there definitely are opportunities there. but you risk stumbling upon having to make performance compromises. 3d noise is expensive 15:52 celeron55 however if you want to go with predictable fractals or honeycomb structures or something, that would be cheap 15:52 Blockhead256 so something like a mandelbulb/julia thereof? 15:52 celeron55 (well, some fractals are expensive also. but i mean, just pick and choose) 15:54 Blockhead256 one could use a single 3d noise and apply a threshold for air/stone and make an interesting shape. Playable? Maybe. 15:56 proller what if make shape of terrain on client with using something like raytrace ? 15:56 celeron55 the way voxel mapgens are made is worth learning for any programmer. it's a serious exercise in advanced pseudorandomness and it could put you into the right mindset even for things like cryptography. i'm not kidding 15:59 Blockhead256 proller: where does the ray come from, where is it pointed, and what does it reflect off? 16:00 proller ray from player, no need to reflect, just touch surface and draw something 16:00 Blockhead256 ok now what algorithm decides what the surface's shape is? 16:00 celeron55 Blockhead256: that's basically how the caves work, except that they use the absolute value of the result and consider values closer to 0 to be air. it creates caves that kind of resemble the interfaces between soap bubbles, the white parts 16:01 celeron55 (the bubble parts aren't round though. they are all kinds of shapes) 16:01 celeron55 (anyway, the insides of the bubbles are stone) 16:03 celeron55 (and, there's not one 3d noise, but two, which are made to interact with each other, to make it more interesting) 16:03 Blockhead256 yeah I went and washed my hands so I could visualise it and that makes sense 16:05 celeron55 anyway, looking at the code, i think we may not be getting our money's worth out of the noise operations. the interaction is just a multiplication. using some kind of interpolated lookup into a curve of some sort could work better, or something else 16:05 Blockhead256 the fractal mapgen just does mandelbrot set and julia sets right? 16:05 proller Blockhead256, server can send mapgen params to client, client can run dedicated mapgen with one function has_surface_at(v3) 16:06 celeron55 src/mapgen/cavegen.cpp is where it's at, if you want to play with it 16:06 Blockhead256 It would be cool to have fractal flames. I have a library of flames as my desktop backgrounds. Not sure how exactly to turn them into terrain though 16:08 Blockhead256 proller: Sounds like the occlusion culling of blocks that the server does before sending them to the client, except instead of sending to the everything to the client you would just ask the client to render anything that's visible but not sent as 'different'. 16:08 Blockhead256 anyway a lot of server operators don't want to send the mapgen parameters, they even ask for ways to hide the seed and keep the noise parameter tweaks to themselves 16:10 proller its only for very far areas for drawing with very low res 16:10 celeron55 that is a good point, the noise params can spoil the world 16:10 celeron55 i don't recall anyone bringing that up ever 16:10 proller i have some working code with drawing cloud of far surfce points 16:10 proller with any mapgen 16:11 Blockhead256 pro tip: If you don't want anybody to be able to replicate your mapgen, give them the real seed as usual but change the seeds of your noise params to be different from default and it should turn out completely different anyway 16:12 celeron55 the seed offsets of the noise params essentially work like a password salt there 16:13 celeron55 of course, giving the seed is meaningless then 16:14 celeron55 you might as well give a random unrelated number 16:14 Blockhead256 I was actually thinking one should send a random number every time the seed is sent to a client if you really wanted to obscure it. Not really much different from a fake seed, but then you can even get people arguing about what the "real" seed is 16:16 Blockhead256 "CavesRandomWalk .. is very fast, executing in less than 1ms on average .. on a modern processor" <- this comment ages well only because processors get faster over time 16:19 proller https://youtu.be/SgwVzX-Czv0 16:21 Abiblima back 16:21 Abiblima oh im Abiblama lol 16:23 Blockhead256 Abiblama: yep that's you :) 16:23 Blockhead256 proller: very cool. Actually it reminds me of 80s sci-fi, just put it on a night background and make it glow 16:24 proller https://www.youtube.com/watch?v=mG_IY17S4_Q 16:25 proller just draw some mesh from this cloud of points and enjoy full view range on any non lua mapgen 16:28 Blockhead256 ok, so the "secret" is how has_surface_at(v3) is implemented. Menger sponge is easy to pre-program. For existing mapgens, does it sample the perlin noise? 16:29 Abiblima wow thats mad bird bones are extended lungs 16:30 proller Blockhead256, its not a secret, its just call to existent mapgen code 16:31 Blockhead256 Interesting. Does it take much to calculate? Does this show that saving the nodes to disk is part of what takes so long? 16:35 proller not much, but need to make many rays, not need to make it every frame, only when you moves to long distances 16:35 proller its not saving mapgen, just calculates points 16:35 Blockhead256 it seems like it would use more memory too 16:36 Blockhead256 understandable that longer view distance does that.. 16:38 Abiblima why would people want to hide their seed lol is it that bad if someone copied it? 16:39 Blockhead256 some people are really fragile about people wanting to replicate their server, however noteable or not 16:39 Blockhead256 it *is* really easy to do with world downloading + seeds + the amount of code that's open source on most Minetest servers 16:40 Blockhead256 There's no way Minecraft would *ever* build world downloading into the game unlike Minetest 16:40 calcul0n_ and as said earlier the seeds can spoil the world, ie you can guess where ores or rare items will spawn 16:41 MTDiscord hiding can make sense for example with mods that use the seed to decide where to spawn stuff 16:41 Blockhead256 I mean you could just x-ray. There's a million ways to cheat, the only thing stopping you is self control most of the time 16:41 proller you can use patched client to view all ores and other hidden things 16:41 proller and server owner can disable sending map params to client if too paranoid 16:42 Blockhead256 it's only a small problem in PvE/survival if someone ruins their own experience by cheating. In PvP it's kind of important, yeah. 16:42 Abiblima oo lol yeah know where all the good ores are ag 16:43 Abiblima teardown 16:43 Abiblima taredown* 16:43 Blockhead256 You can world download and zoom around in singleplayer in all cases. See the contents of locked chests, etc. You can use a cheat client, and so on. 16:44 Blockhead256 the real problem that the average person using the official client has though is that most of us can't stop ourselves from pressing F5 to see our coords all the time 16:44 Abiblima yeah i guess for your average go stealing the seed is a technique anyone can do 16:44 Abiblima jo* 16:45 Abiblima whats wrong with seeing ur coorf? 16:45 Abiblima voxel games are so doomed 16:46 Abiblima will we ever get a better voxel engine than teardown 16:46 Abiblima which is not even that good :( 16:46 Blockhead256 Most times in multiplayer, when you wanna meet your friend you just tell each other your coords and start walking. Or you use /spawn. Or you just walk towards each others' nameplates. No challenge. 16:47 Abiblima yeah 16:47 Abiblima well how else would you meet your friend lol 16:47 Blockhead256 that's fine if it's just a chatroom. If it's about "realism" then you can see some big problems 16:47 Abiblima ah yeah 16:47 Abiblima tbh minetest is just a chatroom 16:47 Abiblima but users can place and dig on the side 16:47 Abiblima but its mostly used a chatroom 16:47 Blockhead256 so accurate... 17:01 celeron55 it's how most online games end up being, at least creative ones 17:02 celeron55 some servers are basically literal virtual shopping malls full of bored teenagers 17:03 celeron55 at least at times 18:26 lissobone wow celeron 18:27 lissobone What did I skip??? 18:27 lissobone Why is celeron55 HIMSELF here??!! 18:28 celeron55 i'm always lurking. the challenge is to say something interesting enough so that i comment about it 18:28 ROllerozxa O.O 18:35 Abiblima he's just slow to respond, still stuck on an old cpu die 18:36 Abiblima celeron55: why 55? 18:37 Abiblima see what i mean he's still computing the answer 18:38 celeron55 it would be 550, but back in the day irc had a maximum nickname length of 9 characters 18:39 Abiblima why did you like that processor at the time or something? 18:39 celeron55 imagine that, ROllerozx 18:39 Ryzen7200 im a better version of you 18:41 lissobone I have discovered a new hobby: bullwhips. 18:41 celeron55 i was just a kid and it was the best i could get. my dad could only get me a 200MHz pentium II which wasn't that great 18:41 Ryzen2700 how old are ya now? 18:42 Ryzen2700 so you went from pentium 2 to celeron 550? what year was this? 18:42 lissobone Probably year 55 BC. 18:42 Ryzen2700 is this before feet pics took over the internet and only fans? 18:42 celeron55 somewhere in the mid 2000s 18:42 Ryzen2700 makes sense 18:42 Ryzen2700 i bet your rich now huh 18:42 Ryzen2700 finnish, feet pics, only fans 18:43 celeron55 oh you're not on my onlyfans? 18:43 Ryzen2700 are you offering a discount on the first month? 18:44 celeron55 no, i'm not cheap like that 18:44 celeron55 i've got to respect myself 18:44 Ryzen2700 h-h-hawt :hotemoji: 18:45 Ryzen2700 early 2000's habbo hotel and runescape 18:45 Ryzen2700 neopets and uh gangwars? 18:45 Ryzen2700 newgrounds :S 18:45 celeron55 i ran an apache web server on windows 2000s 18:45 celeron55 -s 18:45 celeron55 it was wild 18:46 Ryzen2700 lol wow people claim me/2000 was the best release 18:46 Ryzen2700 technically ME and 2000 a different releases right? 18:46 celeron55 me was crap, 2000 was nice 18:46 Ryzen2700 2000 was the good one then lol, what was wrong with me? 18:47 Ryzen2700 i used 2000 for a long time until the directx situation forced me to xp 18:47 celeron55 of course, if you wanted to run dos games 2000 was a bad idea and 98se was the thing 18:47 Ryzen2700 ah lol 18:47 Ryzen2700 and now we have dosbox 18:47 Ryzen2700 when did you quit windows for linux 18:48 Ryzen2700 and what exactly did u host on that apache webserver 18:49 ROllerozxa I don't know what I should be surprised by more, the fact IRC nicknames could only be 9 characters long maximum or that c55 got an onlyfans? 😳 18:49 celeron55 i'd have to look that up from somewhere 18:49 celeron55 ROllerozxa: i'll spoil that for you: only one of those is a fact 18:52 ROllerozxa hmm, well 9 characters *did* seem like an odd amount to limit it at... 18:56 proller celeron55, maybe its time to merge https://github.com/minetest/minetest/pull/12142 ? 18:58 Ryzen2700 how old does the mt version have to be 18:58 Ryzen2700 i've never found a server too old to connect to? 19:01 Ryzen2700 i really liked the vehicles in teardown 19:12 Pexin Ryzen2700: before winXP, there was the win95+ line and the NT line. "millenium edition" was 9x line, 2000 was NT. NT used more resources because better memory segregationand other stability/security features, so required (at the time) more ram. win2k was good, winME was pretty much "why bother?" since there was (to my knowlege) no big improvement over 98 19:12 Pexin then XP came and said everything is NT now because 9x/ME sucks. which wasn't wrong 19:13 Pexin also XP is when mandatory online activation happened. you could install and use 2k whenever you wanted 19:13 celeron55 winxp came and made 2000 look like a kid's toy 19:13 celeron55 i mean, xp is really just a stylized 2000 19:13 celeron55 of course it changed after the service packs 19:14 Pexin xp also had a lot of built in apps, which was ...controversial 19:14 celeron55 2000 was good until ms stopped security updates 19:14 celeron55 lol i really worded that first line badly. i mean, originally xp seemed like just 2000 with a new toyish style 19:15 Pexin I pretty much skipped xp. used 2k long past its lifespan, then went to ..which version of linux did I start on again.. 19:15 celeron55 that's a good question for me also 19:16 Pexin oh mandrake 19:17 celeron55 that's what i used alongside 2000 19:17 Pexin well I mean yeah initiall I dualbooted while I learned it 19:18 celeron55 i'm trying to remember what the distro might have been when i stopped booting to 2000, but i have no idea. i guess i had some games there that i didn't want to delete but didn't actually use. i mean, i have that on my current system too, i have no idea how many years it has been since i booted windows on this 19:20 theyre16charlong I've used Windows XP for the majority of my life, and I saw my father use it when I was very little. 19:21 Pexin right XP got so popular that ms was essentially forced to continue supporting it long after its planned EOL to try avoiding a windows virus pandemic due to consumers not upgrading 19:22 lissobone My dad even bought a licensed disk for ♂Three Hundred Bucks♂. Now he considers that a mistake. It's still (unfortunately) installed on the remaining old computer in our house. I've never seen older M$ operating systems in real life. 19:23 lissobone I discovered something newer in 2016 when my parents decided to buy me a thick samsung laptop. 19:23 lissobone It was windows 7. 19:23 celeron55 i can find some of my own screenshots from 2007, with programs like kde, opera and amarok in them, but these give me no clue what the distro is 19:24 celeron55 i wish kde 3 was still a thing 19:25 lissobone It were the greatest of times, and it were the worst of times. I made many friends while playing throughout all those years on some minecraft server, and I lost all of them. 19:25 celeron55 it was so ahead of its time it's still ahead of everything the wm/de people are doing 19:26 lissobone KDE is nice. But my primary operating system now is GNU Emacs, so I don't bother what window manager am I running. 19:27 celeron55 ha, now i found a screenshot with emerge running in one of the konsoles. so this one is gentoo 19:27 lissobone I have got a cool screenshot of my KDE setup with 100500 transluscent widgets (and a photo of...) 19:27 celeron55 gentoo with kde 3, what a cool kid 19:27 lissobone (A photo of...) (I forgot the name of that bird.) 19:28 celeron55 kde then and kde now are totally different things 19:28 celeron55 same thing for gnome 19:28 celeron55 back then these things catered mainly for power users 19:29 celeron55 these days they cater for clumsy touchscreens 19:29 celeron55 the world has truly changed 19:29 celeron55 for the worse 19:29 Pexin xfce4 19:36 Ryzen2700 yeah it was brutal when they end service packs for a windows release 19:37 Ryzen2700 lissobone: ur missing out windows 3.x 19:37 celeron55 here's one of my screenshots from 2007 https://i.imgur.com/jr6A6qP.png 19:39 celeron55 that laptop died within 3 years from that screenshot, then i got the cf-51 on which i started minetest in 2010 19:39 celeron55 that's a hp pavilion dv2000 series. nice machine, but very well known for dying in a couple of years from new due to the gpu chip desoldering itself 19:40 Ryzen2700 I remember that UI from the old knoppix install cd's 19:40 Ryzen2700 K Desktop Environment 3 is the third series of releases of the K Desktop Environment. There are six major releases in this series. After the release of KDE 4, version 3.5 was forked into the Trinity Desktop Environment. 19:40 Ryzen2700 ^quote from google 19:40 Ryzen2700 what did u like about it so much? 19:43 celeron55 kde 3 was more than a sum of its parts. it's difficult to explain. it was a surprisingly coherent whole, with lots of options and widgets out of the box, which all worked, and still it managed to conserve pixels and resources relatively well. it felt like a product, not like a project 19:44 celeron55 it had themes that were easy to customize. these days customization tends to mean "you can select a pre-made theme". in kde 3, it meant "you can pick any color and size you want for any component in the base theme that you choose" 19:46 Ryzen2700 thats cool, i never got into linux back then, too me years to finally give up on windows 19:46 celeron55 both for the wm stuff and for the toolkit itself 19:46 celeron55 and it had all those options in a good ui. same thing for widgets, choose any you want, and that's not it, they all had lots of options 19:47 Ryzen2700 i might boot it in a vm sometime 19:47 celeron55 kde 4 was an absolute trainwreck when it came out and i realized kde 3 was going to be gone in less than a year or so 19:47 Ryzen2700 if the iso is still easy to find 19:48 celeron55 i had to switch to something else, and i think i settled on icewm, which i still use. it's a wm stuck in the 90s, and i love it for that. no updates to mess up my experience since 20 years ago 19:49 Ryzen2700 https://www.reddit.com/r/kde/comments/vp77h5/it_might_be_an_unpopular_opinion_but_kde_4_was/?rdt=50543 19:49 Ryzen2700 nice im on xfce 19:49 Ryzen2700 its awesome 19:49 Ryzen2700 uv happy 19:49 Ryzen2700 v* 19:50 celeron55 kde 4 was creative, and that was about it. it wasted so many pixels, was resource hungry, the apps like amarok were redesigned to be absolutely useless 19:50 celeron55 they seemed to drop all the hard work they had done do kde 3, in exchange for... something i didn't understand at all 19:51 celeron55 s/done do/done for/ 19:51 celeron55 it was traumatizing, really 19:51 celeron55 i probably have PTSD about kde 4 19:52 Ryzen2700 lmao 19:52 Ryzen2700 like blender 7 to 8 19:52 celeron55 it's not something i wish for, or would allow, anyone to experience 19:52 Ryzen2700 sounds like your a little autistic buddy 19:53 Ryzen2700 icewm, is that linked to icecat browser 19:53 Ryzen2700 is there like a whole ice ecosystem 19:53 Pexin how about when I abandoned gedit for pluma? 19:53 celeron55 icewm also is not something i wish anyone to experience, but it's the only thing that is keeping me sane in terms of window managers 19:53 Ryzen2700 does icewm look like a windows clone? 19:54 Ryzen2700 lol i ditched pluma for mousepad 19:55 Ryzen2700 now im just ditched mousepad for pluma last week 19:55 Ryzen2700 tbh im now considering setting files to open with nano with mouse pointer enabled 19:55 Ryzen2700 using xfce-konsol 19:55 Ryzen2700 or gnome-console 19:56 Ryzen2700 gnome-console is a little better in ways i am aware of 19:56 celeron55 i use pluma when i need a graphical editor instead of vim which i usually use in the terminal 19:57 Pexin I use nano unless I'll be switching between documents a lot, then pluma 19:57 celeron55 and for the terminal, i use urxvt 19:57 celeron55 this is what my icewm looks like https://i.imgur.com/DoMMNTc.png 19:57 Pexin never wrapped my brain around vim 19:57 celeron55 basically just a classic unix X11 wm look 19:58 Ryzen2700 nice 19:59 Pexin you got an irssi plugin to colorize usernames? 19:59 Ryzen2700 you a rightie? 19:59 Ryzen2700 rightie controls 19:59 Ryzen2700 shocking 20:00 celeron55 this theme is based on the motif window manager which is from 1989. i modified it to have the bright yellow title for the focused window so that i never miss what's in focus 20:00 g_ woops 20:00 Ryzen3700 dammnnn 20:00 Ryzen3700 killed the window accidentally 20:00 Ryzen3700 but still 20:00 celeron55 rightie window controls? 20:01 Ryzen3700 min max and close buttons lol 20:01 celeron55 of course, i grew up on windows 20:01 Ryzen3700 same but i flipped for linux 20:01 Ryzen3700 and never looked back 20:01 Pexin isn't that a mac thing? 20:01 Pexin xfce defaults to right side 20:02 Ryzen3700 well i started on ubuntu 16.04 20:02 celeron55 in icewm you can of course choose how you place the window title buttons, any way you like 20:02 Ryzen3700 was it ubuquity wm back then or something 20:02 Ryzen3700 unity? 20:03 Ryzen3700 yes celeron but the point is you are so windows stockholm that your still suffering the traumatic conditioning to this very day 20:03 Ryzen3700 saying that you're not far from stockholm are you 20:03 Ryzen3700 conspiracy much? 20:04 celeron55 oh you're a psychologist now? nice 20:04 Ryzen3700 lol i gttos go sleeps now ill let you know in morn'' 20:04 lissobone it's 5 am for me lol 20:04 celeron55 i need to sleep too 20:05 lissobone i can't sleep because i am making ropes out of strings 20:05 lissobone help me i am now attached to little strings that control me 20:05 Pexin https://www.youtube.com/watch?v=iAykOz1gWi4 20:06 lissobone i think im going insane with handcrafts 20:06 * celeron55 imagines pulling from the rope, which makes lissobone spin like a lawnmower engine and then lissobone starts up 20:08 lissobone there's no more room for starting up, all my engines are already at top power (unfortunately for both my sleep and body temperature regulation cuz it's hot outside) 20:08 lissobone like in that popeye scene when he ate spinach and his muscles turned into hydro power plants 20:10 lissobone aside from bullwhips and ropes, an idea of a word generator has struck me 20:10 lissobone it will generate words not yet known by humanity 20:10 lissobone in ANY language! 20:11 lissobone i already know how to implement this in emacs lisp 20:12 lissobone i will store the word generation data in an alist, with keys being word lengths and data being the probability tables 20:12 lissobone i can omit probabilities for pure randomness