Time Nick Message 00:04 VanessaE ok, something's just not right here. the server's spiking to 100% just to load terrain that's been generated ages ago 00:10 RealBadAngel hmmmm, no matter the values i try to put there it always crashes with same error 00:13 hmmmm yes i know 00:13 hmmmm gosh i guess you were just lucky with some parameter 00:13 hmmmm it looks like the ore sheet too needs that check before i call pr.range() 00:15 RealBadAngel so what in the example is wrong? 00:15 hmmmm nothing, it's my mistake 00:16 hmmmm but for some reason i feel like the ore it's producing after being fixed isn't enough 00:16 hmmmm lemme try it in the air 00:16 RealBadAngel ok 00:17 hmmmm ah nevermind there's quite a bit 00:18 hmmmm it's just not what you'd expect underground 00:18 hmmmm alright i'll push a fix 00:18 hmmmm i still wonder why i didn't see that bug earlier 00:18 hmmmm must be soemething about those parameters 00:20 hmmmm https://github.com/minetest/minetest/commit/8d13967aa3a802df8fa8d97fcce32fea2b46f2bf 00:22 RealBadAngel testing 00:40 realbadangel_ hmmm no luck at all 00:40 hmmmm what? 00:41 realbadangel_ im flyin underground and cant find any 00:41 hmmmm keep looking, you'll see it 00:41 hmmmm when you find some there will be a lot of it 00:42 hmmmm and that's the whole point of this ore type, is it not? 00:42 realbadangel_ yeah 00:42 realbadangel_ what is the way to make it more common? 00:42 realbadangel_ i mean which setting 00:43 hmmmm it is already very common 00:43 hmmmm you'll see if you were to make it replace air and have it above ground 00:43 realbadangel_ going to try 00:44 hmmmm you don't see much of it because it's very short vertically 00:44 hmmmm if you increase the height you should see more, but... 00:51 realbadangel_ now i can see them 00:53 realbadangel_ how to adjust the size of it? 00:54 realbadangel_ spread? 00:59 RealBadAngel looks like 00:59 RealBadAngel hmmm, noticed one thing 01:00 RealBadAngel they do generate in pattern, only at certain height levels 01:02 RealBadAngel http://i.imgur.com/ya5WYGN.jpg 01:03 RealBadAngel http://pastebin.com/9UJK2MKL 01:03 hmmmm ? 01:03 hmmmm what would you prefer it to do? 01:03 RealBadAngel see the picture? 01:03 hmmmm yea? 01:04 RealBadAngel they do spawn in layers. between layers there are no generated sheets 01:04 hmmmm isn't that how it's supposed to spawn...? 01:04 RealBadAngel its like 10 nodes in height between each layer 01:04 hmmmm yea 01:05 hmmmm at least 10 nodes in height 01:05 hmmmm usually more 01:05 RealBadAngel was guessing 01:05 hmmmm so are you saying that you don't want it to do this? 01:05 RealBadAngel i would expect it the height should be more random 01:06 RealBadAngel because if i define range for them like 100 nodes in height 01:07 RealBadAngel there will be heights in that range that sheet will never be spawned 01:07 RealBadAngel and this is the same for whole world 01:08 hmmmm so are you saying that you'd like more than one to spawn per chunk? 01:08 RealBadAngel no, they should spawn on different heights in given range 01:08 RealBadAngel now they do spawn always at the same height 01:09 hmmmm oh i see what you're saying 01:09 hmmmm crap 01:09 RealBadAngel 10 nodes above, another layer 01:09 RealBadAngel +20 another 01:09 RealBadAngel nothing in between 01:09 hmmmm that shouldn't happen 01:09 RealBadAngel pick my example and see it for yourself 01:09 hmmmm i'm passing along the blockseed 01:10 hmmmm yeah that's not happening, you just saw it in one chunk 01:20 RealBadAngel no, i can see it all around 01:20 hmmmm all the ore in the same chunk is at the same height 01:20 RealBadAngel http://i.imgur.com/lLKcHRV.jpg 01:20 RealBadAngel look, its bigger than one chunk. dont you think? 01:20 hmmmm well one chunk is 80x80x80, so 01:20 RealBadAngel always same heights 01:20 RealBadAngel well, maybe in 80x its true 01:20 RealBadAngel but anyway it just doesnt look any random :) 01:20 hmmmm you're free to improve it 01:20 RealBadAngel maybe if sheet could go up or down a bit? 01:20 RealBadAngel like for given diameter few nodes height change allowed 01:20 hmmmm it does - it's completely random https://github.com/minetest/minetest/blob/master/src/mapgen.cpp#L155 01:20 RealBadAngel height is random, but sheet is flat 01:20 RealBadAngel i mean to twist whole sheet 01:20 hmmmm oh.. 01:20 RealBadAngel so it could go up or down 01:20 hmmmm well gosh, you don't really want a sheet, do you. 01:20 RealBadAngel twisted sheet :) 01:20 hmmmm =/ 01:20 hmmmm try a scatter ore with 3d noise and very high concentration (low scarcity value) 01:21 RealBadAngel but then i will get it everywhere 01:21 hmmmm looks like you should invent a new ore type then 01:21 RealBadAngel i will play with settings 01:22 RealBadAngel maybe i will invent somethin 01:22 hmmmm well 01:43 RealBadAngel hmmmm, hell yeah :) 01:44 RealBadAngel its way better imho 01:46 hmmmm mmmm 01:46 hmmmm alright, i'll add that 01:46 hmmmm but what i have there wasn't really the intent of the sheet ore type 01:47 RealBadAngel so maybe call it another type? 01:47 hmmmm no, because all i did was add a line of code and change another 01:47 hmmmm alright 01:48 RealBadAngel its still a sheet, but looks way way better 01:48 hmmmm here's how i'm going to do this: you need to modify the scale of the noiseparams to get how bendy it is 01:48 hmmmm but 01:48 hmmmm you need to disregard the scale when you go to write the threshhold value 01:49 RealBadAngel so flat one will be still possible, yes? 01:49 hmmmm yes, by leaving the scale at 1 01:49 hmmmm or 0 perhaps 01:50 RealBadAngel thats good 02:04 RealBadAngel hmmmm, you got it ready? 02:38 hmmmm yeah 02:38 hmmmm https://github.com/minetest/minetest/commit/26cad481d08432df1c9ace5f1ac23f1fad9a403f 02:39 hmmmm it's a shame this goes underground 02:39 hmmmm it generates really neat looking things 02:39 VanessaE hmmmm: you never know what modders will do to abuse the API hooks they're given 02:40 hmmmm yeah, with the same ore generation type you can already make nice looking clouds 02:40 hmmmm now with this you can do pretty nice floating islands 02:40 VanessaE I was *just* thinking about that 02:42 RealBadAngel or cover ground level with something 02:43 hmmmm haha oh wow 02:43 hmmmm that's true, you could make lakes or springs with this 02:43 hmmmm which would naturally create rivers 02:43 hmmmm ore = water_source, wherein = dirt_with_grass 02:44 hmmmm and you'd need a very scarce noiseparams 02:45 RealBadAngel i think this one is gonna have many uses :) 02:45 VanessaE hmmmm: oooooo 02:46 VanessaE people HAVE been asking when we'll have proper rivers..... 02:46 hmmmm this isn't a proper river though 02:46 VanessaE yeah Iknow 02:46 VanessaE but it's the closest thing yet 03:07 hmmmm well... this isn't quite what i was expecting.. 03:07 VanessaE ? 03:08 hmmmm http://ompldr.org/vaHZpYg 03:09 VanessaE err 03:10 VanessaE back to the drawing board..... 03:10 VanessaE :_) 03:10 hmmmm http://ompldr.org/vaHZpYw 03:10 VanessaE :) 03:13 Exio just some tweaks and profit! :D 03:21 Exio going to sleep now, bye! 03:43 RealBadAngel hmmmm, what were the settings for this? http://ompldr.org/vaHZodQ 03:47 hmmmm http://pastebin.com/tJ3fWNsq 03:47 hmmmm whoops 03:47 hmmmm oh no that's right 03:58 RealBadAngel i like it 03:59 RealBadAngel bbl, work time 04:02 hmmmm i think i figured out what is it about prestidigitator that bothers me 04:02 VanessaE which would be? 04:02 hmmmm he doesn't work with us at all, so there's really no input and we can't guide him or comment on anything, and he's going to show up with this huge patch and expect us to just plop it in 04:04 hmmmm also he thinks he's better than us because he strictly follows moronic super-object-oriented coding guidelines that belong in a java enterprise application and not minetest 04:04 VanessaE actually the first part of that is sort of how I saw it 04:04 VanessaE he takes his own direction, etc. 04:06 hmmmm unfortunately for his super-duper polymorphic interface shitfuck noise class with inheritance, that "10x slower" single-point noise penalty is more than enough to make it completely unappealing 04:07 VanessaE right, and we don't need that 04:07 hmmmm but we wouldn't know what he's doing wrong because we haven't seen anything of his 04:07 VanessaE you did a hell of a lot of work to speed the engine back up to 0.3.x levels (probably beyond?) 04:07 hmmmm we still extensively use single-point noise, so it's necessary 04:07 hmmmm i thought 0.3.x was slower because of the 3d noise 04:08 VanessaE well a lot of folks have said that 0.4.x is slow by comparison - or was. I've seen the same too, but it ain't anymore 04:08 VanessaE but you know, that stuff's pretty subjective 04:08 hmmmm maybe it is faster, definitely not graphics-wise 04:08 VanessaE you've got benchmarks to prove it ;-) 04:09 hmmmm the most recent minetest is practically unplayable on my laptop 04:09 hmmmm come on, i thought minetest was designed for crappy graphics cards 04:09 VanessaE well that much is true, but since RBA's "6d" patch, I've gained back most of what I lost in graphic performance 04:10 hmmmm yeah 04:10 VanessaE you saw my comment earlier about 512px mode working again? 04:10 hmmmm i haven't, but i believe you 04:11 hmmmm before i continue onto the next big thing, i think i want to identify what's causing the stutering 04:11 hmmmm i have a rough idea of where to look, it's somewhere in between 0.4.3 and 0.4.4-d1 07:26 celeron55 hmmmm: shaders? 07:26 celeron55 or the switch to using opengl as default on windows 07:26 celeron55 those seem like the uttermost likeliest causes 08:08 celeron55 lol, the nodebox is still marked experimental in lua_api.txt 8) 10:42 Exio celeron55: because it is! 14:21 RealBadAngel hi all 14:22 RealBadAngel first of all we shall take care of memory management 14:22 RealBadAngel when game starts. we go back to menu 14:23 RealBadAngel and after a few times doing so it crashes 14:23 RealBadAngel we got problem with destructors 15:35 PilzAdam the favorite server list is broken in win builds 15:35 PilzAdam the GUI doesnt show any servers and client/serverlist/favoriteservers.txt only contains the most recent server 15:43 hmmmm mmmmmmmmmm, proller__ 15:44 proller__ mee 15:44 hmmmm thexyz_ we need the dev wiki captcha fixed 15:48 proller__ maybe setup recapcha ? 15:48 hmmmm yeah, anything that actually works is better than this cat stuff. 15:49 proller__ PilzAdam, maybe my bug 15:52 proller__ PilzAdam, but i dont change this logic, it seems broken when i make online serverlist 15:53 proller__ and it not working on unix with not-root user 16:15 thexyz_ hmmmm: sure 16:15 thexyz_ should work now 16:26 PilzAdam I dont like that the public server list is shown when switching to the Multiplayer tab 16:31 proller__ maybe save last choice ? 16:32 PilzAdam yep 16:33 proller__ but local list is broken and not usable 16:33 PilzAdam it works for me 16:34 proller__ i make fix for saving later 16:49 VanessaE local list doesn't work for me, either 16:49 VanessaE only the public list does. 16:50 VanessaE (linux, self-built from git) 16:55 rubenwardy hi all 16:55 PilzAdam thexyz_, I am reading a bit through the code about the nick completion 16:56 PilzAdam seems like the clientenvironment only has the local player 16:56 PilzAdam and the server doesnt send anything to the client to add new clients 16:57 PilzAdam I found TOCLIENT_PLAYERINFO that sends the player names, but its marked as obsolete 17:10 PilzAdam the players are only send as active objects to the client 17:15 VanessaE couldn't you just extract the info you need from those "*** so and so left/joined" messages? 17:15 VanessaE it's hacky but it wouldn't require any new network overhead 17:17 hmmmm hmmm it seems that the Minecraft equivalent to our EmergeManager is called "WorldManager" 17:18 hmmmm except it's much more tightly coupled with their mapgen, making the whole setup much less flexible 17:18 Exio minecraft flexible? 17:35 proller__ PilzAdam, please add desc to wiki for indev mapgen: * Added a new map generator called ''indev'' (rare HUGE caves, float lands at y=500+, higher mountains and biome sizes at edges) 17:35 proller__ i cant select cars 8) 17:35 proller__ cats 17:36 Exio cats? what? 17:36 proller__ http://dev.minetest.net/index.php?title=Changelog&action=submit 17:36 proller__ cats 8( bit i prefer dogs 17:36 hmmmm you don't need to do that anymore 17:36 hmmmm it was removed 17:37 Jordach why do mountains have to be higher at the edges? 17:37 Jordach then it seems like your in a crater 17:37 proller__ for more intresting exploring 17:38 proller__ and more harder 17:38 proller__ no, sea deeper too at edges 17:38 proller__ try it 17:39 proller__ at server DEV TESTING MAY WIPE! 17:39 proller__ can teleport 17:41 proller__ and! all far-depend scales adjustable via mapgen params 17:45 proller__ you can set 0.001 and have flat at edges 18:16 PilzAdam workaround for nick completion: https://github.com/PilzAdam/minetest/commit/38a6a7f5af3572d3100701f1d342e3a9238cce7b 18:21 PilzAdam it inserts the player name to a list in the clientEnvironment when the playerCAO is generated 19:08 hmmmm PilzAdam, you're familiar with the entity system, yes? 19:08 hmmmm i want to add the persistent entity attribute 19:09 PilzAdam in Lua? 19:09 hmmmm that is, if false, the entity will not be saved on chunk unload 19:09 hmmmm obviously this must be done in the core 19:09 PilzAdam I only know the Lua side of entities well 19:09 hmmmm oh, okay. 19:10 hmmmm to your knowledge, is there any sort of setting that's like this in any way? 19:10 Exio there was a patch for adding force_field or so 19:10 hmmmm that's different 19:10 Exio maybe you can get "info" in how you can add the settings or stuff like that to it 19:10 Exio ah 19:19 PilzAdam hmmmm, maybe in environment.cpp:1718 19:21 hmmmm there already is something like this, okay... why is it that i keep hearing about /clearobjects from users and how it takes such a long time 19:21 hmmmm i thought that they needed to do that because objects kept falling and they piled up so they had to clear them eventually 19:22 hmmmm i hope dropped nodes, saplings, etc. are all objects that get deleted on unload. 19:22 Exio what about a *timed* remove? 19:23 hmmmm that's pretty wasteful because you'd have to track each object 19:23 PilzAdam hmmmm, if the mapbock unloads the objects just convert into a static form 19:23 PilzAdam /clearobjects deletes static objects too 19:24 hmmmm it gets deleted if m_known_by_count == 0 19:27 hmmmm hmm okay 19:28 hmmmm so how do we do this so /clearobjects isn't necessary very often 19:29 PilzAdam block objects at the map edges, add item decaying 19:29 hmmmm i would prefer if only objects with a certain attribute ever become static in the first place 19:30 hmmmm if we were to add item decay that'd require a timer for each object 19:31 proller thexyz_, Please select all the cat photos: 8( 19:33 hmmmm so that's just one part of the equation 19:34 hmmmm the other part is to make clearobjects not murder peoples' ram 19:35 hmmmm i am thinking of doing this by having loadBlock() to increment a "# blocks loaded" statistic to the Map whenever a block is inserted into a MapSector 19:36 hmmmm if it passes a threshhold in EmergeThread, or emergeBlock, it'll unload the least recently used block (they're sorted by expiry time already) 19:36 hmmmm yes? no? 19:39 PilzAdam how does that help? 19:39 hmmmm by not eating peoples ram by having 1000000000000000000000000 blocks loaded at the same time 19:41 PilzAdam why are the blocks loaded when the objects are removed? 19:41 PilzAdam isnt there another way to drop the static objects? 19:41 hmmmm of course not 19:41 hmmmm static objects are stored in the blocks on disk 19:42 hmmmm clearobjects literally loads the entire map into ram 19:45 proller can we optimize block loading-generation when fallin down? 19:45 proller to load bottom only at first 19:47 Calinou how are you supposed to fall then 19:47 Calinou hello, opensuse user! 20:53 proller for falling without stops, or minimize stops 20:54 proller if you know speed and direction - you can load [first|only] on way blocks 20:55 hmmmm indeed, the first blocks to be sent should be those in the immediate 6 neighbors 20:56 hmmmm in your case it's probably network latency that's the reason why you 'stop' - the block loading itself is decently fast 20:59 PilzAdam it isnt network latency, it even happens in singleplayer 21:00 hmmmm i'm just saying the problem is definitely not on the emergethread side at all 21:02 proller but when your speed 100+ per second - you dont need in 6 blocks, only your and next by way 21:03 hmmmm obviously that's an extreme case and isn't common at all 21:03 hmmmm minecraft sidesteps this problem completely because they load in whole vertical sectors 21:04 proller minecraft is 2d 21:05 hmmmm i'm just saying, maybe it isn't too horrible if there is a delay 21:05 proller yes, but better without delay 8) 21:06 PilzAdam maybe add a max vertical velocity for the player 21:06 proller on thexyz_s server i found 4000+ handmade hole 21:06 proller PilzAdam, nooo, its very cool to fall on biiig speed 21:06 hmmmm when i say "maybe it isn't too horrible", i mean "this is not and won't be a priority for me at all" 21:07 PilzAdam you cant reach high speeds because you hit a unloaded block 21:07 proller yes, but you can jump twice 21:07 Calinou how' bout making players stuck but keep their velocity on unloaded blocks 21:07 Calinou back in september 2011 or so this was the case 21:07 Calinou (0.2.20110922_3) 21:08 hmmmm if that were so, you'd be able to fall out of the map 21:08 Calinou just stuck ;) 21:08 Calinou we need some kind of instantly-killing lava at layer -30000 21:08 Calinou or just bedrock :p 21:09 VanessaE "hard iron" 21:09 VanessaE use that new layered ore generation to do it :-) 21:10 VanessaE wait, it *could* do that, couldn't it? 21:10 hmmmm it could, but that would be a waste 21:11 VanessaE true 21:11 hmmmm ugh, i don't know what i want to accomplish next 21:11 hmmmm fix clearobjects, redo block sending, finish mapgen v7 21:12 VanessaE I could make various random suggestions, but all would be somewhat less than trivial :-0 21:12 VanessaE :-) 21:12 proller another idea - -31001 threat as +3999 21:12 proller 39999 21:12 proller ups 21:12 VanessaE proller: hell no 21:12 proller 30999 21:13 proller on any direction 21:13 VanessaE now, X/Z directions that makes good sense 21:13 VanessaE but I doubt it's doable 21:13 proller why not y? 21:14 VanessaE because you fall down, why should you do anything other than eventhally land on something? 21:14 hmmmm that would fit in with the whole retro theme that mine* games have 21:14 PilzAdam proller, I teleported to y=10000, then fell down to 0 (with breaks), then teleported back to y=10000 so I can fall without interuptions 21:14 VanessaE hmmmm: which would? 21:14 hmmmm y wraparound 21:14 VanessaE oh 21:14 PilzAdam at 0 I toogled free move, and stopped at -6328 21:15 hmmmm i suppose you could do mg_name = singlenode with mapgen_singlenode == air 21:15 hmmmm and test it that way 21:16 VanessaE PilzAdam: which is busted all to hell IMHO. 21:16 proller PilzAdam, yes, i tried this 21:16 VanessaE I *hate* the new falling/climbing model 21:16 proller but without free move toggle 21:18 PilzAdam hmmmm, I even hit unloaded blocks with singlenode mapgen 21:23 VanessaE back when we had minetest.ru:30001 it was very difficult to hit an unloaded block while falling 21:24 Exio in what hardware did it run? 21:24 VanessaE well, it was "not common" anyway 21:24 VanessaE dunno, ask thexyz_ 21:24 Exio falling 21:25 Exio i'm able to get unloaded chunks for a very few ms every 5~6 seconds 21:25 Exio with singlenode and air 21:27 Exio ahahahaha 21:27 Exio rofl, i just fell... in a nyancat! 21:28 VanessaE heh 21:31 proller http://servers.minetest.net/ : i want to remove :30000 default port if its default and place servers without passwords first, any objections? 21:31 proller for simpler join for new players 21:32 VanessaE no, don't do that 21:32 VanessaE passwords should be encouraged 21:33 proller but its too hard to get passwort to random server 21:34 VanessaE get? 21:34 VanessaE no 21:34 VanessaE you just have to make one up. 21:34 VanessaE 123456 is better than no password at all 21:34 proller ups 21:35 proller agree 21:37 celeron55 the way minetest handles passwords tends to be a bit hard to grasp to first-time users 8) 21:38 celeron55 i've never seen any other software do it the same way 21:38 rarkenin Hashing and base64? 21:38 celeron55 lol no 21:38 celeron55 the UX 21:38 VanessaE celeron55: ironic, considering that it really can't be any easier :-) 21:38 celeron55 = no registration at all 21:38 celeron55 VanessaE: yes 21:38 proller on some server you cant join without registration, but we havent flag about this 21:39 rarkenin Why not have each server generate 4096-bit PGP keys, then send the password encrypted that way? 21:39 rarkenin Oh, saw your previous message. That last one about PGP was not meant to actually be implemented. 21:39 celeron55 rarkenin: yeah sure, add more build dependencies for windows 21:47 celeron55 feel free to search for lightweight cryptography libraries though... something to do an unauthenticated key exchange and something to do some AES encryption, probably 21:47 rarkenin_ OK. I'll look. Perhaps some sort of lightweight RSA? 21:53 PilzAdam should I add sound_place to Item or NodeDef? 21:54 celeron55 here are some medium-sized ones that can probably be taken seriously: http://botan.randombit.net/ http://libtom.org/ http://beecrypt.sourceforge.net/ 21:54 celeron55 i've been using crypto++ at work, but it's huge 21:54 Kray hey, don't you have some semiofficial server listing thingie now! 21:55 celeron55 Kray: we have an official automatically working one 21:55 Kray just grab public key from that and then you can be sure the connection is secure 8D 21:56 rarkenin_ Possibly. 21:56 Kray (that would be overkill) 21:56 Kray you are only exchanging unimportant passwords 22:04 RealBadAngel hi all 22:05 RealBadAngel a proposition: since we can play local sounds (to client only) without download content 22:05 RealBadAngel couldnt we make music for client? 22:08 Exio i'd like that 22:08 Exio (Ambience?) 22:08 RealBadAngel this could be server independent 22:09 RealBadAngel c++ code can play local sounds (somebody was testing it and left there working code) 22:10 RealBadAngel i already made for me technic ambience with script on the server side but music local 22:13 RealBadAngel code is a bit of test one, it is limited to just 10 sounds stored loclally with fixed names+number 22:13 RealBadAngel but its easy to fix 22:14 RealBadAngel i tested it with movin all ambience/technic ambience tracks to the client side 22:14 RealBadAngel it works perfect 22:34 Exio i'd like some sort of client-side scripting 22:35 rarkenin Seconded. 22:35 Exio just limiting features because "cheaters may use them" seems just a soft limit 22:35 Exio and even, the client-side scripting should not need to have access to a lot of stuff 22:35 rarkenin And/or running server-side mods on clients for performance. Depending on server settings, the mods could also be run server-side either fulltime to prevent cheating, or just at randon times for spot-checking. 22:36 rarkenin *random 22:36 Exio that is something what i'd like too, but it seems "too much work" :P 22:36 rarkenin Also data structures need to be shared. 23:01 PilzAdam different place sounds for nodes: https://github.com/PilzAdam/minetest/commits/sound2 23:16 PilzAdam any comments? 23:18 Exio if i understad the code correctly, the sounds "aren't" confirmed by the server? (my main worry is the lag) 23:18 ShadowNinja The place sound is the grass footstep sound 23:19 PilzAdam ShadowNinja, thats just a testing sound in minimal 23:19 PilzAdam Exio, yes, its client side like the other sounds 23:20 Exio awesome 23:20 Exio grr, understand* i need to write better :P 23:37 hmmmm Kray, that's what I said, but celeron thinks that because people tend to use the same passwords for everything, the password on a minetest server is a valuable target (which I disagree with.. my password on remote servers is asdf, and I bet it's similar for nearly everybody else) 23:38 Exio now i can hack your accountz! 23:38 * Exio uses H4CKT3STPR0.EXE 23:39 hmmmm well, I guess the new challenge is to identify my accounts 23:40 Exio i can guess them: one random key/char + other key repeated N times 23:40 Exio :P 23:40 hmmmm so you believe my nick is the result of randomness, not thinking? 23:41 hmmmm one day i was thinking of what my nick should be for freenode, so i put "hmmmmm..." 23:41 Exio you just think-ed in random stuff 23:43 hmmmm anyway when i just woke up before i had a new idea 23:44 hmmmm isn't determining the terrain characteristics based on biome the sameish as modulating the noise values by another noise? 23:44 hmmmm i determine the biome by noise, so... 23:44 hmmmm what i do is have that modulating perlin 2d map vary the characteristics of the base terrain noise map 23:45 hmmmm then what i do is go through the terrain i just generated again and determine what biome it is by the characteristics of the terrain and the biome of nearby terrain 23:45 hmmmm which is, erm, how it's done in real life.