| --- Log opened Sat Sep 03 00:00:25 2011 |
| 00:02 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 00:05 | < ToxicFrog> | That actually strikes me as quite unexpected. |
| 00:05 | < ToxicFrog> | I'd expect either $PATH or $LD_LIBRARY_PATH, depending on OS. |
| 00:06 | < ToxicFrog> | (and $PYTHONPATH for the modules themselves) |
| 00:09 | | Derakon [chriswei@Nightstar-f68d7eb4.ca.comcast.net] has quit [[NS] Quit: Lost terminal] |
| 00:09 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 00:17 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 00:38 | | Derakon[AFK] is now known as Derakon |
| 00:52 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 00:58 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 01:25 | | Phox [Phox@Nightstar-3e5deec3.gv.shawcable.net] has quit [Ping timeout: 121 seconds] |
| 01:27 | | LANrah [Z@26ECB6.A4B64C.298B52.D80DA0] has quit [[NS] Quit: If the world didn't suck, we'd all fall off.] |
| 01:54 | | Phox [Phox@Nightstar-3e5deec3.gv.shawcable.net] has joined #code |
| 02:20 | | Derakon is now known as Derakon[AFK] |
| 02:53 | | Kindamoody[zZz] is now known as Kindamoody |
| 03:10 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has joined #code |
| 03:17 | | * Vornicus argls. WRITE CODE DAMN YOUR EYES |
| 03:18 | <@McMartin> | Hm. I should really fire up Game Maker this weekend. |
| 03:18 | <@McMartin> | Finish up the main menu UI. |
| 03:43 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 04:13 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has quit [Ping timeout: 121 seconds] |
| 04:42 | < Rhamphoryncus> | So I had an insane thought earlier |
| 04:42 | < Rhamphoryncus> | I keep wanting a map tile system that can rotate in smaller increments, ie rather than 90? (square) or 60? (hex) it'd do 30? or something |
| 04:43 | < Rhamphoryncus> | But that's impossible. All the angled distances are irrational, no way to make them line up exactly |
| 04:44 | < Rhamphoryncus> | But the catch: that's assuming euclidean space. |
| 04:44 | < Vornicus> | You might be better served with a square or hex thing but with larger minimum steps. |
| 04:45 | < Rhamphoryncus> | Rendering to 3d already involves various kinds of distortion, so could I use some subtly non-euclidean space and have the difference be invisible when rendered? |
| 04:45 | < Rhamphoryncus> | Vornicus: That's the closest you can do normally, but I want to setup a template of buildings, then rotate and place it and have it Just Work |
| 04:47 | < Vornicus> | THere is as far as I am aware actually 0 distortion when rendering 3d on a planar thing. |
| 04:48 | < Rhamphoryncus> | Depends how you define distortion :P |
| 04:48 | < Rhamphoryncus> | fov shows distortion to me |
| 04:49 | < Rhamphoryncus> | You change your angle and things warp a bit. You don't notice because it's entirely natural |
| 04:50 | < Vornicus> | Welll... |
| 04:50 | < Vornicus> | You could render your world on a poincare disc. |
| 04:50 | < Vornicus> | And only show a very small portion of that disc. |
| 04:50 | < Rhamphoryncus> | I have no idea what that is, which makes it awesome :D |
| 04:50 | | * Rhamphoryncus googles |
| 04:51 | < Vornicus> | A poincare disc renders a hyperbolic plane onto a euclidean plane by turning straight lines into circles that meet perpendicularly with a single outer circle that shows up at infinity. |
| 04:51 | < Rhamphoryncus> | I really need more math knowledge :( |
| 04:52 | < Rhamphoryncus> | So far it hurts my head, but it seems like just the thing I mean by non-euclidean geometry |
| 04:52 | < Vornicus> | The difficulty is that you're now trading distortion for size. |
| 04:53 | < Rhamphoryncus> | mmm, squares not using right angles |
| 04:54 | < Vornicus> | http://en.wikipedia.org/wiki/Uniform_tilings_in_hyperbolic_plane |
| 04:55 | < Vornicus> | Check out particularly the order-5 square tiling. But honestly if you want even semi-arbitrary rotation you're stuck with real-world. |
| 04:55 | < Rhamphoryncus> | floating point. I hate floating point :( |
| 04:56 | < ToxicFrog> | Why? |
| 04:57 | < Rhamphoryncus> | So it's possible to create the tiling, the question is if you can tweak it such that you don't notice you're not in normal euclidean space.. |
| 04:57 | < Rhamphoryncus> | It's fuzzy |
| 04:57 | < ToxicFrog> | Only in certain well-defined circumstances. |
| 04:58 | < Rhamphoryncus> | rotating would give me a distance that's an irrational number. Does it collide with another object at the edge or not? Fuzzy. |
| 04:59 | < Rhamphoryncus> | The best solution I have to that fuzziness is to make collision fuzzy to. Let buildings be squished slightly at the cost of extra build time or something |
| 04:59 | < Vornicus> | Not that fuzzy. |
| 05:00 | < Vornicus> | Seriously, it's nowhere /near/ that fuzzy. |
| 05:00 | < Rhamphoryncus> | Only slightly fuzzy is still fuzzy. Drives me nuts |
| 05:00 | < Rhamphoryncus> | Particularly if you're trying to fit buildings tightly |
| 05:00 | < Vornicus> | Let me describe to you precisely how fuzzy. |
| 05:01 | < Rhamphoryncus> | You may build a template in one way, then rotate it and have them fail a fitness check |
| 05:01 | < Vornicus> | You have really screwed up fitness checks then. |
| 05:01 | < Rhamphoryncus> | Or heck, you could just change unrelated code, rebuild, and have the same template fail in the original position |
| 05:01 | < Rhamphoryncus> | No, you just didn't add a tolerance for the fuzziness |
| 05:02 | < Vornicus> | Look, it's like this. the fuzziness is on the order of 1 nanometer over the distance of the continent. |
| 05:02 | < Vornicus> | If you're complaining about fuzziness you're doing it wrong. |
| 05:03 | < Rhamphoryncus> | You basically need to have the editor place at one distance while the sanity check uses a slightly smaller distance. |
| 05:03 | < Rhamphoryncus> | eh I understand how it works |
| 05:03 | < Vornicus> | ANd you're complaining about something that happens with integer coordinates too. |
| 05:04 | < Vornicus> | you've seen those "missing square" puzzles? |
| 05:04 | < Rhamphoryncus> | no |
| 05:04 | < Rhamphoryncus> | hrm maybe |
| 05:04 | < Vornicus> | http://en.wikipedia.org/wiki/Missing_square |
| 05:04 | < Vornicus> | I can generate these with as close an angle as you please. |
| 05:06 | < Rhamphoryncus> | Those don't matter. I'm only talking about placing buildings and such. I can limit the user to certain angles, I just want more angles than a square grid |
| 05:11 | < Vornicus> | DO you want the player to be able to walk into the building? Do you expect the exterior and interior to be continuous? |
| 05:13 | < Rhamphoryncus> | I'm thinking more of a strategy type game, but it shouldn't be noticable that it's non-euclidean |
| 05:13 | < Vornicus> | Okay, so, the way TOtal Annihilation does it |
| 05:14 | < Rhamphoryncus> | units move arbitrarily, buildings are on tiles? |
| 05:14 | < Vornicus> | YOur building has a grid size and a model size; the model is free to rotate within the grid it's been allocated so long as it doesn't have to extend beyond its grid to get there. |
| 05:21 | < Vornicus> | (You see this while playing as the fact that you can build things based on tiles, but then when you select them they have a slightly smaller, and rotated, box around them.) |
| 05:21 | < Rhamphoryncus> | I get the impression people don't care much about tightly placing buildings |
| 05:22 | < Vornicus> | This happens even when the buildings are "tightly placed" |
| 05:27 | < Rhamphoryncus> | this a fair screenshot? http://cataclysm.tauniverse.com/CORE.gif |
| 05:27 | < Vornicus> | Yes. Look at the fusion plants to the right. |
| 05:28 | < Vornicus> | Those have been tightly placed, and their model borders do not include the "stairs" |
| 05:28 | < Vornicus> | Notice how the arms of the various plants are at different angles. |
| 05:29 | < Rhamphoryncus> | Looks like they use a square bounding box aligned on the grid, then rotate the model based on its center |
| 05:29 | < Vornicus> | Yep. |
| 05:29 | < Vornicus> | THe model has a slightly smaller "true" bounding box. |
| 05:29 | < Rhamphoryncus> | I respect that design. It's *sane* |
| 05:30 | < Rhamphoryncus> | I want better though ;) |
| 05:30 | < Vornicus> | TOugh shit :) |
| 05:30 | < Rhamphoryncus> | heh |
| 05:30 | < Rhamphoryncus> | It's an academic exercise anyway |
| 05:33 | < Rhamphoryncus> | Another approach may be to have both local and global grids. Global could be triangular (simplest possible), local changes to match your building orientation but has to add up to the same total area similar to the missing square problem |
| 05:34 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has joined #code |
| 05:35 | <@Tarinaky> | Wow. The graphics of TA weren't nearly as good as I remembered them. |
| 05:36 | < Rhamphoryncus> | lol |
| 05:36 | <@Tarinaky> | Screenshot looks hideous. |
| 05:36 | <@Tarinaky> | >.< |
| 05:38 | < Vornicus> | Tarinaky: quite a few of those models don't exist in the original game, including the shield bubbles, the giant aircraft, the giant gun, the dragons' teeth, and the large fusion plants. |
| 05:38 | < Vornicus> | I'm not sure what's up with most of the fusion plants; the ones that don't have the obvious grid are what they look like in the original game. |
| 05:38 | < Vornicus> | It was 1997; software rendering was the only way to go. |
| 05:39 | < Vornicus> | WHich means no normals, minimal at best lighting, and so forth. |
| 05:42 | <@Tarinaky> | Dragon's Teeth were totally part of one of the official patches. |
| 05:43 | < Vornicus> | Not /those/ dragon's teeth. |
| 05:43 | <@Tarinaky> | Oh. |
| 05:43 | < Vornicus> | THough I guess those might be the walls from CC. They sucked, I never used them. |
| 05:44 | < Rhamphoryncus> | yeah, that's a mod, but it does show the grid |
| 05:44 | < Vornicus> | But they were very ugly, because they tried to use a single model for what any sane thing would have made multiple models. |
| 05:44 | < Rhamphoryncus> | which thing are you talking about? |
| 05:44 | < Vornicus> | THe walls, seen around the big big gun. |
| 05:44 | < Rhamphoryncus> | ahhh |
| 05:45 | < Rhamphoryncus> | Which brings up another point: it's often better to place the building wherever, then alter the rendering to fit the space |
| 05:45 | < ToxicFrog> | I don't think those are TACC walls; they were 1x3 or 1x4 sheets of metal, IIRC. |
| 05:47 | < Rhamphoryncus> | Another way to think of this whole problem: could you use a triangular grid and NOT be annoyed by it? |
| 05:48 | < Vornicus> | No. |
| 05:48 | < Vornicus> | Triangle grids are a royal pain in the ass; they have 9 connections, only three of which are edges. |
| 05:48 | < Rhamphoryncus> | so? |
| 05:49 | < Vornicus> | So how many different directions do you have keys for naturally on the numpad? |
| 05:49 | < Rhamphoryncus> | I mean as a player in a strategy game like that |
| 05:49 | < Rhamphoryncus> | Scrolling doesn't match the grid anyway |
| 05:52 | < Vornicus> | So I don't get to see the grid at all? |
| 05:53 | | celticminstrel [celticminstre@Nightstar-5d22ab1d.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 05:57 | < Rhamphoryncus> | You see it when you place buildings, sure |
| 05:57 | < Rhamphoryncus> | are you thinking you'd use the keypad when placing buildings? |
| 06:01 | < Vornicus> | I was wondering what things would see the triangle grid. Other thing about the triangle grid is that things don't really feel tightly packed with it. |
| 06:02 | < Vornicus> | if you have 1-triangle objects, in order to fit in the triangles they have to be small enough that there's room for one more in the middle. |
| 06:03 | < Rhamphoryncus> | That's how I end up feeling with a square grid. I'm very conscious that it's a game mechanic |
| 06:03 | < Vornicus> | A hex grid, you still get the 30deg rotations and it packs tighter. |
| 06:04 | < Rhamphoryncus> | You can have most buildings be hexes even if the underlying grid is triangular |
| 06:06 | < Vornicus> | Not what I mean |
| 06:06 | < Vornicus> | A true hexagonal grid, even the smallest possible building, packs buildings in without large enough gaps for buildings of the same size. |
| 06:08 | < Vornicus> | Squares also can do that. Triangles, however, one-tile objects have to be so small that you can fit more of those objects on top of the corners. |
| 06:11 | < Rhamphoryncus> | You're talking about the visual aspect of being able to rotate the building? |
| 06:11 | < Vornicus> | No. |
| 06:12 | < Vornicus> | I'm talking "if the building is to fit tightly in the triangle, the building must be this small" |
| 06:13 | < Rhamphoryncus> | A square building? Or..? |
| 06:13 | < Vornicus> | Any building that has a shape that isn't the triangle. |
| 06:13 | < Vornicus> | Circular, square... |
| 06:14 | < Rhamphoryncus> | That can be avoided by making the grid half the size. Only trivial buildings (light posts?) would be truly 1 tile |
| 06:20 | < Vornicus> | Or you can use the hex grid, which allows you to still use 30deg rotations, and has the same orientation for every tile, and packs tight, and you can use keys to move around if you want. |
| 06:20 | <@Tarinaky> | May I ask... why're you using a grid? |
| 06:21 | <@Tarinaky> | Wouldn't a quadtree work better? |
| 06:21 | <@Tarinaky> | I guess it'd still, technically, be a (square) grid but it'd be sufficiently fine as to be invisible. |
| 06:22 | < Rhamphoryncus> | Tarinaky: floating point ate muh baybee! |
| 06:23 | <@Tarinaky> | Floating point is more precise if you bound it within 0->1 |
| 06:23 | <@Tarinaky> | And you only, really, need an approximation. The human eye only has so much definition. |
| 06:24 | <@Tarinaky> | Oh FFS. |
| 06:24 | <@Tarinaky> | I don't know how to turn my parents alarm clock off. |
| 06:24 | | * Tamber hands Tarinaky a mallet. |
| 06:25 | < Rhamphoryncus> | Just unplug it |
| 06:25 | <@Tarinaky> | I can't reach the plug. |
| 06:25 | <@Tarinaky> | The plug is behind a nightstand that is stacked sillyhigh. |
| 06:25 | <@Tarinaky> | I pushed a button and it stopped. |
| 06:25 | < Rhamphoryncus> | Even if it's like my alarm clock, with a battery backup, the battery would have died after only a couple months of use. And no, that doesn't include any time being without power and actually using the battery. |
| 06:26 | < Vornicus> | So make your quadtree with like 24 levels and then you have millimeter resolution over an area the size of 2 Disney Worlds. |
| 06:26 | <@Tarinaky> | ^ |
| 06:27 | < Rhamphoryncus> | heh |
| 06:28 | < Vornicus> | 16 levels gets you 1-meter resolution over Los Angeles. |
| 06:29 | <@Tarinaky> | And it's not like they're expensive. |
| 06:29 | <@Tarinaky> | Pointer dereferencing is meant to be one of the cheapest operations or something. |
| 06:30 | <@McMartin> | Not... really so much. |
| 06:30 | < Vornicus> | No. |
| 06:30 | <@Tarinaky> | Oh, okay. |
| 06:30 | <@Tarinaky> | >.> <.< >.. |
| 06:30 | <@McMartin> | A pointer deref goes out to memory, or at least to cache |
| 06:30 | < Vornicus> | THat. |
| 06:30 | <@McMartin> | Arithmetic operations never leave the registers. |
| 06:31 | <@Tarinaky> | I did say one of >.> |
| 06:31 | < Vornicus> | Hah, not even that. |
| 06:31 | | * Vornicus gives Tarinaky a nanosecond. |
| 06:31 | < Vornicus> | (it's a foot of wire) |
| 06:31 | < Rhamphoryncus> | I find myself looking at this tiling: http://en.wikipedia.org/wiki/File:Tile_V46b.svg |
| 06:32 | < Vornicus> | .../eight/ orientations for triangles. |
| 06:32 | < Vornicus> | Ew, no. |
| 06:32 | <@Tarinaky> | At a nanosecond you can do it 100,000 times without even profiling it. |
| 06:33 | <@Tarinaky> | Or something like that. |
| 06:33 | < Rhamphoryncus> | That lets me have rectangular buildings with 12 orientations |
| 06:33 | <@Tarinaky> | Probably even more. |
| 06:33 | < Vornicus> | ANd you have to do millions of them per second. |
| 06:33 | <@Tarinaky> | That's still only 1ms. |
| 06:34 | < Vornicus> | See, that wire you hold in your hand |
| 06:34 | < Vornicus> | that's one way. |
| 06:34 | <@Tarinaky> | IDGI. |
| 06:35 | < Vornicus> | You need two of those to deref a pointer. Oh, and it's not just a deref here; if your pointer points at a structure, then you need another half a nanosecond to add in the structure offset for the thing you want. |
| 06:35 | < Vornicus> | And then you've got... |
| 06:35 | < Vornicus> | Here's the thing. Pointer dereferences are cheap, sure |
| 06:35 | | * Tamber tosses "What Every Programmer Should Know About Memory" at Tarinaky? |
| 06:35 | <@Tarinaky> | 60Hz is slightly less than 16ms. |
| 06:35 | < Vornicus> | Compared to, I don't know, the elementary functions. |
| 06:35 | <@Tarinaky> | Err |
| 06:35 | <@Tarinaky> | Slightly more eeven. |
| 06:36 | <@Tarinaky> | Most monitors only do 60Hz so more draws than that is wasted effort. |
| 06:36 | < Vornicus> | But then you're doing millions of them for even the most elementary state updates in the game. |
| --- Log closed Sat Sep 03 06:42:09 2011 |
| --- Log opened Sat Sep 03 06:42:16 2011 |
| 06:42 | | TheWatcher [chris@Nightstar-3762b576.co.uk] has joined #code |
| 06:42 | | Irssi: #code: Total of 22 nicks [7 ops, 0 halfops, 0 voices, 15 normal] |
| 06:42 | <@Tarinaky> | I think I'm in the wrong field :p |
| 06:42 | | Irssi: Join to #code was synced in 38 secs |
| 06:42 | < Rhamphoryncus> | Bwahaha, I think that tiling actually does what I want |
| 06:43 | <@Tamber> | Tarinaky, you think /you're/ in the wrong field; I don't even know whos field I'm in, but there's a bull looking very angrily at me... |
| 06:44 | < Rhamphoryncus> | If you have triangles that are 1 unit long then you end up with rectangles that are 1.5x2.828... |
| 06:44 | < Rhamphoryncus> | So they're irrational in one dimension |
| 06:44 | < Rhamphoryncus> | But that doesn't matter at all. |
| 06:44 | <@Tarinaky> | Tamber: Terrible joke, I wouldn't even laugh if I wasn't so tired :p |
| 06:45 | <@Tamber> | :p |
| 06:45 | < Rhamphoryncus> | Hmm those numbers don't look right |
| 06:46 | | * Tarinaky sets an alarm and gets some kip. |
| 06:46 | < Vornicus> | WHich edge is 1 now? |
| 06:46 | < Rhamphoryncus> | The rectangle I've drawn only has a 1:1.16 ratio |
| 06:47 | < Rhamphoryncus> | Vornicus: the triangle's edge is 1. The rectangle's short edge is 1.5 |
| 06:47 | < Vornicus> | Which. triangle's edge. |
| 06:47 | < Vornicus> | It's got three. |
| 06:47 | < Rhamphoryncus> | Equilateral, sorry |
| 06:47 | < Vornicus> | Okay, then... |
| 06:47 | < Rhamphoryncus> | http://en.wikipedia.org/wiki/File:Tile_V46b.svg |
| 06:48 | < Rhamphoryncus> | Going from the top left and moving right it's only 2 edges |
| 06:48 | < Vornicus> | your rectangle's short edge is sqrt(3)/2, long edge is 3/2. Hexagon circumradius is 2. |
| 06:48 | < Vornicus> | You're using some ridicularge triangles there. |
| 06:49 | < Rhamphoryncus> | Yeah, odd mental space, sorry. I'm going from my previous idea and subdividing it |
| 06:49 | < Rhamphoryncus> | So in my head a "triangle" is the original equilateral triangle |
| 06:49 | < Rhamphoryncus> | and the non-equilateral chunk is.. whatever it is :) |
| 06:50 | | * Alek eyes that tiling. |
| 06:50 | < Alek> | I can actually see it being done with equilaterals. |
| 06:50 | < Alek> | would be a pretty nasty spikefield. |
| 06:50 | < Vornicus> | Actually |
| 06:51 | < Alek> | I'm not sure it'd be on a plane, either. |
| 06:51 | < Alek> | what WOULD happen if you converted those triangles to equilaterals? |
| 06:51 | < Vornicus> | THere are at least three sizes of equilateral triangle smaller than the one you saw. |
| 06:52 | < Vornicus> | Alek: nothing interesting, equilateral triangle tiling is regular. |
| 06:52 | < Vornicus> | But that tiling there |
| 06:52 | < Alek> | no, the way it's tiled there. |
| 06:52 | < Rhamphoryncus> | ergh, I once again confused myself. My rectangles are missing one of the rotations |
| 06:52 | < Alek> | take THAT tiling, then morph all the triangles to equilaterals. |
| 06:53 | < Vornicus> | Rham: the other problem here is that now you've got /three different corners/ |
| 06:55 | < Vornicus> | Alek: yeah, they'd overlap and collapse to regular. |
| 06:56 | < Alek> | um. |
| 06:56 | < Alek> | what if you transformed it into 3D? |
| 06:56 | < Vornicus> | (I can count equilateral triangles with 2, 6, 8, 18, and 24 small right triangles in there) |
| 06:57 | < Vornicus> | ALek: I don't know what you're on about at all now. |
| 06:57 | < Alek> | take that plane. |
| 06:57 | < Alek> | transform all the triangles to equilaterals. |
| 06:57 | < Vornicus> | It won't line up at all in 3d. |
| 06:57 | < Alek> | let the plane deform into a 3D object to fit. |
| 06:57 | < Alek> | no? |
| 06:58 | < Alek> | hm. oh well. |
| 07:05 | < Rhamphoryncus> | So if I did this using rectangles but pretending they're square I'd end up with a 2x5 building being 3.464x7.5 going one way and 3x8.660 the other way |
| 07:05 | < Rhamphoryncus> | 15% distortion |
| 07:06 | < Vornicus> | Which is getting into the noticeable range. |
| 07:06 | < Rhamphoryncus> | Yup |
| 07:06 | < Rhamphoryncus> | But thought of another way, that could be padding around the buildings |
| 07:06 | < Rhamphoryncus> | They wont' fit so tight, but ehh.. |
| 07:07 | < Rhamphoryncus> | Or you could tweak the rendering so it uses the full space |
| 07:07 | < Rhamphoryncus> | It is a rectangular space after all. The edges are flat and conform to your rectangular building |
| 07:08 | < Rhamphoryncus> | Hmm, doesn't let you build a template and rotate it, not unless everything is using these rectangles-as-squares |
| 07:11 | < Rhamphoryncus> | That's fixable if you bump up your building scale (as opposed to layout scale) up a notch, so it's all 2x2 rectangles instead. |
| 07:18 | < Vornicus> | Or, you could just use quadtrees and get over yourself. |
| 07:20 | < Rhamphoryncus> | That hardly achieves my desires |
| 07:30 | < Vornicus> | Too bad. :) |
| 07:30 | < Rhamphoryncus> | By bumping up the ratio enough I can drop the distortion to 1.03% |
| 07:32 | < Rhamphoryncus> | The question at that point: is the difference in placement tile size (the tiny equilateral triangles) vs building tile size (almost square rectangles) a problem? |
| 07:35 | < Vornicus> | Maybe, amybe not. But your placement tiles are all sorts of crazy orientations. |
| 07:36 | < Vornicus> | You've got 12 different orientations for the placement tiles, which is a recipe for a lot of crazy edge cases. |
| 07:36 | < Rhamphoryncus> | Well.. the placement uses the major apexes, ie the equilateral apexes. It's just the occupancy that subdivides further |
| 07:38 | < Vornicus> | There are two different "apexes" for equilateral triangles. |
| 07:39 | < Rhamphoryncus> | ergh I meant corner |
| 07:39 | < Vornicus> | Which one! |
| 07:39 | < Vornicus> | THere's two. |
| 07:39 | < Rhamphoryncus> | you mean 3? |
| 07:39 | | Derakon[AFK] is now known as Derakon |
| 07:39 | < Rhamphoryncus> | heya Derakon |
| 07:40 | < Derakon> | Hullo. |
| 07:40 | < Vornicus> | Two distinct ones. |
| 07:40 | < Rhamphoryncus> | I don't get it |
| 07:40 | < Vornicus> | the 6- and 12- fold symmetry centers. |
| 07:40 | < Vornicus> | Or rather they're not actually symmetry centers. |
| 07:41 | < Vornicus> | But the points with 6 and 12 right triangles around them, respectively. |
| 07:41 | < Rhamphoryncus> | I have no idea what you mean |
| 07:42 | < Vornicus> | A small equilateral triangle has two distinct places where its corners could be. |
| 07:42 | < Vornicus> | Unless of course you mean a different triangle, in which case you should pick one, I can't tell. |
| 07:43 | < Rhamphoryncus> | Are you confused about equilateral? |
| 07:43 | < Vornicus> | No. |
| 07:43 | < Rhamphoryncus> | 3 identical sides |
| 07:43 | < Vornicus> | There's quite a lot of differently-sized equilateral triangles here. |
| 07:44 | < Rhamphoryncus> | the smallest one |
| 07:44 | < Derakon> | Scalene, isosceles, equilateral. Why do I remember that first word? |
| 07:44 | < Rhamphoryncus> | It gets chopped up into 6 non-equilateral triangles |
| 07:45 | < Vornicus> | THat's not the smallest. |
| 07:45 | < Vornicus> | The smallest is 2. |
| 07:45 | < Rhamphoryncus> | oops :/ |
| 07:45 | < Rhamphoryncus> | touch? |
| 07:46 | < Vornicus> | In fact I can see 7 different sizes that all fit into that swatch of pattern there. |
| 07:46 | < Vornicus> | sorry, 9. |
| 07:47 | < Rhamphoryncus> | Right, justified confusion |
| 07:48 | < Rhamphoryncus> | And I got my tile size backwards in gimp, but I can't tell the difference, heh |
| 07:49 | < Vornicus> | heh |
| 07:50 | < Rhamphoryncus> | Actually, it looks all wrong if I do it the "right" way. Clearly I'm doing something wrong |
| 07:50 | < Vornicus> | Note that things won't look quiiiiiite like right angles if you have the ratios off. |
| 07:52 | < Rhamphoryncus> | Anyway, the bottom line here is that by scaling up a triangular grid I can have square-ish tiles that are arbitrarily close to true square |
| 07:53 | < Vornicus> | Except that once again the angles will look off. |
| 07:53 | < Rhamphoryncus> | Hmm? |
| 07:54 | < Rhamphoryncus> | Once you hit 1% distortion it's easier to have the building themselves square and just add some extra grass around them |
| 08:04 | | Derakon is now known as Derakon[AFK] |
| 08:11 | | Vash [Vash@Nightstar-f03c5637.sd.cox.net] has quit [[NS] Quit: <3 I Lovecraft Vorn <3 Nyaaaaaaaaaaaaaaa~ ] |
| 09:41 | | AnnoDomini [annodomini@60F158.737D66.CA1918.CC7562] has joined #code |
| 09:47 | | AnnoDomini [annodomini@60F158.737D66.CA1918.CC7562] has quit [[NS] Quit: leaving] |
| 10:40 | | LANrah [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code |
| 11:30 | | Vornicus is now known as Vornicus-Latens |
| 11:45 | | LANrah [Z@Nightstar-3602cf5a.cust.comxnet.dk] has quit [[NS] Quit: If the world didn't suck, we'd all fall off.] |
| 11:54 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 12:06 | | GeekSoldier [Rob@42F7BC.2D44AA.6F437F.C36BF3] has joined #code |
| 12:06 | | GeekSoldier [Rob@42F7BC.2D44AA.6F437F.C36BF3] has left #code [] |
| 12:32 | | Kindamoody is now known as Kindamoody|out |
| 13:13 | | Kindamoody|out is now known as Kindamoody |
| 13:20 | | Attilla [Some.Dude@Nightstar-f29f718d.cable.virginmedia.com] has joined #code |
| 14:20 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [[NS] Quit: ChatZilla 0.9.87 [Firefox 6.0.1/20110830092941]] |
| 15:24 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 16:32 | | Kindamoody is now known as Kindamoody|out |
| 18:21 | | Rhamphoryncus [rhamph@Nightstar-14eb6405.abhsia.telus.net] has quit [Client exited] |
| 18:21 | | Derakon[AFK] is now known as Derakon |
| 18:28 | | celticminstrel [celticminstre@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 19:01 | | Kindamoody|out is now known as Kindamoody |
| 19:16 | | cpux_ [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 19:16 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 19:17 | | cpux_ is now known as cpux |
| 19:21 | | Kindamoody is now known as Kindamoody[zZz] |
| 20:12 | | Vornicus-Latens is now known as Vornicus |
| 21:46 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [Connection closed] |
| 21:49 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 23:01 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has quit [Client closed the connection] |
| 23:03 | | cpux [chatzilla@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 23:15 | | * Vornicus gets graph building working for basic rooms, at least. |
| 23:34 | | Derakon is now known as Derakon[AFK] |
| 23:52 | <@McMartin> | Maybe I should write my hex picker today. |
| --- Log closed Sun Sep 04 00:00:39 2011 |