| --- Log opened Sun Jun 29 00:00:09 2008 |
| 00:20 | < Vornicus> | Okay, filtering goes like this: I have a full list of all the points that are currently available, and I get three sorts of them, one for each resource. |
| 00:22 | < Vornicus> | Then I take one of the three sorts, iterate over the items in the sort, and partition each sort according to the item, anything at or equal to in that sort is held on to. then I intersect those three partitions, and if that's nonempty, I can filter that item out. |
| 00:22 | < Vornicus> | What this does is it removes any where i can get at least as good in all three by another point. |
| 00:23 | < Vornicus> | The remaining items become the production possibility frontier, and I can work with that alone without having to worry about the filtered out points. |
| 00:23 | < NSGuest-7122> | i did a bubble sort once |
| 00:24 | < Vornicus> | ...the only trick is I have to check for exact duplication. |
| 00:25 | < Vornicus> | My current filter assumes there are no exact duplicates; I don't know how to prevent it from removing both. |
| 00:26 | < Vornicus> | ...wait, I know. I have to remove from all three sorts when I filter, then when the second one comes along it goes "okay, I don't see anything else here, so I don't have to filter. |
| 00:26 | < Vornicus> | ...yeah. |
| 00:39 | | gnolam [lenin@Nightstar-19083.NA.cust.bahnhof.se] has quit [Quit: Z?] |
| 00:52 | | Zombina is now known as AnnoDomini |
| 00:57 | < Vornicus> | hm. Okay, to get full expressiveness for the output I need to calculate the threshold for each planet to use the planet buff facilities. |
| 01:55 | | * Vornicus manages that. |
| 02:22 | | * Vornicus convinces it to produce a list of production possibilities. |
| 02:24 | < Vornicus> | 300 for one planet. THis is the biggest type... smaller planets are more common. Thank goodness, because this is about to hurt. |
| 02:35 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Quit: <Insert Humorous and/or serious exit message here>] |
| 02:42 | | AnnoDomini [AnnoDomini@Nightstar-28972.neoplus.adsl.tpnet.pl] has quit [Quit: Nobody makes jokes in base 13.] |
| 03:08 | | * Vornicus examines his code. I'm actually surprised at how simple that was. |
| 03:11 | < Vornicus> | Probably still not fast though. |
| 03:12 | < Shoukanjuu> | There is nothing more hardcore than olympic fencers. |
| 03:21 | < Vornicus> | good god, that took an age. |
| 03:23 | < Vornicus> | the good news on the other hand is that it rejected... 3/4 of the available frontier points. |
| 03:34 | < Vornicus> | So, yeah. Really cool, but it takes an age. |
| 03:35 | < NSGuest-7122> | they found the guy who started the "obama is a muslim" smear |
| 03:35 | < NSGuest-7122> | what are some creative revenge schemes having to do with email chain letters |
| 03:35 | | NSGuest-7122 is now known as Doctor_Nick |
| 04:06 | | * Vornicus fiddles with progressive reporting... |
| 05:37 | | * Vornicus tries to figure out how to make a "glom all these together using binary elimination" thing. |
| 05:39 | | * Vornicus tries to figure out how to make it stop... |
| 05:54 | < Vornicus> | there. |
| 05:54 | < Vornicus> | okay, what other tools do I need. |
| 06:47 | < Vornicus> | Need to actually deserialize the data. That'd help. |
| 07:20 | < Vornicus> | it occurs to me that in the end it might be easier to do some sort of algebraic thing - instead of figuring out each individual possible data point, figure the planes... |
| 07:22 | < Vornicus> | ...only problem is I end up with each plane having some nasty constraints.... I don't have any idea how to keep the constraints sane while pushing the planes out. |
| 07:25 | < Vornicus> | Fortunately, there is /an entire field of study/ that covers this. |
| 07:26 | <@Raif> | Vorn: What's the problem you're trying to solve? |
| 07:26 | < Vornicus> | Raif: ever played Space Empires? |
| 07:27 | < Vornicus> | (it's okay to say no, I can explain) |
| 07:28 | <@Raif> | Nope. :) |
| 07:28 | < Vornicus> | Okay. |
| 07:28 | < Vornicus> | in Space Empires, planets are organized into systems, and each planet has a maximum number of facilities. |
| 07:31 | < Vornicus> | There are also three resources, which you gain by having particular facilities. Each such facility produces 1000 * planet value * miscellaneous bonuses: planet value is different for each resource on each planet, and ranges from 0 to 1.5. miscellaneous bonuses are 1.0 + population bonuses (up to 0.9) + happiness bonuses (can be assumed to be 0.2) + racial bonuses (0.2 for minerals, 0 for the otehr two) + facility bonuses. |
| 07:32 | < Vornicus> | The number of facilities per planet varies; because of other things I want my planets to have, the number of available facilities ranges from 1 to 23. |
| 07:32 | < Vornicus> | Now. |
| 07:33 | < Vornicus> | I wish to produce the three resources in a particular ratio, to support the maximum number of a particular type of ship. |
| 07:35 | < Vornicus> | In addition to the three resource facilities, there is one planet-level improvement facility per resource which adds a 0.45 to the miscellaneous bonus for extraction facilities on the planet... and one system-level improvement facility per resource which adds a 0.45 to the miscellaneous bonus for extraction facilities of that type. |
| 07:36 | < Vornicus> | and finally there is a "space port", of which there /must/ be at least one in the system, or the system will not produce at all. |
| 07:37 | < Vornicus> | So what I'm trying to do, is figure out the maximum. This involves arranging production and improvement facilities accordingly. |
| 07:39 | < Vornicus> | currently I have some crazy, grindy stuff where I figure all the possible locations in the production space for each planet and then pile all those together to get the complete possibility space (and removing options that are completely behind the possibility frontier as often as possible) |
| 07:41 | < Vornicus> | Now I'm looking at it and saying "wait a second, I can create the planes for each planet and add them together and get the same thing, except with less shrieking insanity." |
| 07:45 | < Vornicus> | Which as I'm thinking about it is becoming slowly more and more shriekingly insane. |
| 07:53 | | ErikMesoy|sleep is now known as ErikMesoy |
| 08:05 | < Vornicus> | So now I'm looking at the data and trying to figure out if I can do this a better way. |
| 08:45 | | Reiver [~reaverta@Admin.Nightstar.Net] has quit [Ping Timeout] |
| 08:47 | | DiceBot [~Reiver@Nightstar-6717.xdsl.xnet.co.nz] has quit [Ping Timeout] |
| 08:51 | | DiceBot [~Reiver@Nightstar-6717.xdsl.xnet.co.nz] has joined #Code |
| 08:52 | | Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code |
| 08:52 | | mode/#code [+o Reiver] by ChanServ |
| 08:56 | | Vornicus is now known as Vornicus-Latens |
| 10:08 | | AnnoDomini [AnnoDomini@Nightstar-29230.neoplus.adsl.tpnet.pl] has joined #Code |
| 10:08 | | mode/#code [+o AnnoDomini] by ChanServ |
| 11:52 | | Attilla [~The.Attil@92.9.59.ns-21252] has joined #code |
| 11:52 | | mode/#code [+o Attilla] by ChanServ |
| 12:05 | < Doctor_Nick> | what |
| 13:05 | | * McMartin drops an anvil on Doctor_Nick. |
| 13:06 | <@AnnoDomini> | Stop, drop and roll. |
| 13:06 | | * ErikMesoy replaces the anvil with a very small $LANGUAGE script to save Doctor_Nick. |
| 13:06 | <@McMartin> | So, if you hold still, the anvil drops and then the giant boulder runs you down? |
| 13:32 | | Attilla_ [~The.Attil@92.9.59.ns-21252] has joined #code |
| 13:34 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Ping Timeout] |
| 13:35 | < Doctor_Nick> | not an anvil :( |
| 13:44 | | Attilla_ is now known as Attilla |
| 13:44 | | mode/#code [+o Attilla] by ChanServ |
| 14:47 | | gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has joined #Code |
| 14:47 | | mode/#code [+o gnolam] by ChanServ |
| 15:59 | | Attilla_ [~The.Attil@92.9.59.ns-21252] has joined #code |
| 16:00 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Ping Timeout] |
| 16:16 | <@Reiver> | So, uh. Anyone present have moderately up-to-date guidelines on what counts as Decent Hardware these days? |
| 16:17 | | * Reiver needs to build his parents a new computer; knows jack shit about stuff that isn't a year and a half old. |
| 16:19 | | ErikMesoy [~ejm@Nightstar-878.bb.online.no] has quit [Quit: I demand that Reiver go to bed!] |
| 16:22 | | Shoukanjuu [~Shoukanju@Nightstar-20375.dhcp.embarqhsd.net] has quit [Quit: Life is a game.. The object is to survive, to imprint in history your legacy. Go forth and show them what you are, and why you, your memory, and your soul will not be so easily vanquished!] |
| 16:24 | | Shoukanjuu [~Shoukanju@Nightstar-20375.dhcp.embarqhsd.net] has joined #code |
| 17:01 | | Attilla_ is now known as Attilla |
| 19:27 | | ErikMesoy [~ejm@Nightstar-878.bb.online.no] has joined #code |
| 19:27 | < ErikMesoy> | Is TF here? |
| 19:27 | < ErikMesoy> | I just thought of a horrible, horrible pun that I had to share with him. |
| 19:39 | < Shoukanjuu> | share with everyone else then? |
| 19:40 | < ErikMesoy> | After five days of downloading his TA torrent I finally figured out why it's so slow - glacially slow, in fact. It's tied to the rate of climate change! After all, the current ETA for it to finish is the day after tomorrow. |
| 19:40 | | * ErikMesoy flrrd. |
| 19:40 | < Shoukanjuu> | XD |
| 19:41 | < ErikMesoy> | (That was bad, right?) |
| 19:41 | < Shoukanjuu> | iy was |
| 19:51 | <@gnolam> | Reiver: how decent does it have to be? |
| 19:51 | < ErikMesoy> | Seeing as it's 7AM at Reiv's place and he was still up at 3AM, I think he's asleep. |
| 19:53 | < Shoukanjuu> | So uh |
| 19:53 | < Shoukanjuu> | Diablo 3 |
| 19:53 | < ErikMesoy> | Yes. |
| 19:53 | < Shoukanjuu> | \o/ |
| 19:53 | < ErikMesoy> | \m/_O_\m/ |
| 19:54 | < Shoukanjuu> | ?\O_o/? |
| 19:54 | < ErikMesoy> | orz |
| 19:55 | < Shoukanjuu> | People are upset that there is a rainbow in their diablo. |
| 20:01 | < Shoukanjuu> | have you seen the cloth physics |
| 20:01 | < Shoukanjuu> | I find myself excited more about the cloth physics than anything else >_> |
| 20:05 | | * Vornicus-Latens wonders how far in/out you can zoom. |
| 20:12 | < Shoukanjuu> | It must be at least a little..Even so, you can still see the small cloth and things around moving in th gameplay video |
| 20:12 | < Shoukanjuu> | I want CAPES. If only for mopre customization. |
| 20:12 | | * Vornicus-Latens watches Shou get sucked into the demonic equivalent of a jet engine. |
| 20:12 | < ErikMesoy> | I don't think anyone flies in Diablo. |
| 20:12 | < Vornicus-Latens> | tough |
| 20:12 | < ErikMesoy> | The Barbarian in D2 can jump 15 meters, though. |
| 20:13 | < Shoukanjuu> | >_> |
| 20:13 | < Shoukanjuu> | Why does everyone automatically assume that capes are for flying |
| 20:13 | < Shoukanjuu> | I have a cape >:O |
| 20:13 | < Shoukanjuu> | I can't fly |
| 20:13 | < Shoukanjuu> | I can glide with all the style of a brick, though |
| 20:18 | < Vornicus-Latens> | okay. Linear programming in this case is not going to help as much as I thought; the constraint zone is not convex. |
| 20:18 | <@gnolam> | That would complicate things. |
| 20:19 | < Vornicus-Latens> | Quite. |
| 20:19 | | * gnolam shudders. LP. |
| 20:21 | < Shoukanjuu> | I need to document these little things you say |
| 20:21 | < Vornicus-Latens> | Which, mine, or gnolam's? |
| 20:22 | <@gnolam> | (Ok, LP isn't so bad if you have tools adequate for the task *cough*AMPL*cough*, but when you only have a rather crippled MATLAB - yech!) |
| 20:23 | < Shoukanjuu> | Yours and just about anyone else who know swhat it going on |
| 20:23 | < Shoukanjuu> | I imagine a sherlock holmes sort of thing, with Vorn sitting there with his pipe: "....the constraint zone is not convex." And Watson, "That would complicate things." |
| 20:24 | < Shoukanjuu> | I'm still waiting for the "elementary, my dear gnolam" |
| 20:24 | < Shoukanjuu> | Perhaps it is my lack of sleep |
| 20:24 | < Shoukanjuu> | Or a deeper misunderstanding that I have at present for the current state of things |
| 20:24 | < Vornicus-Latens> | On the other hand... I might possibly be able to, uh, how do I put it. |
| 20:26 | < Vornicus-Latens> | If I instead create many different constraint polyhedra and solve each one... |
| 20:26 | <@gnolam> | Shoukanjuu: in very /very/ simplified terms, most LP problems can be thought of as walking around the different corners of a polytope and finding the one that gives you the best result. And if that polytope's not convex, you can "get stuck" in a corner, and fail to reach the optimal solution. |
| 20:27 | < Shoukanjuu> | No, I didn't misunderstand that |
| 20:28 | < Shoukanjuu> | It was a misunderstanding of the type of people and how they would react, or something to that effect |
| 20:28 | < Shoukanjuu> | I'm losing myself in my own words. Dearest me. |
| 20:28 | < Vornicus-Latens> | in this case gnolam is the expert |
| 20:28 | < Vornicus-Latens> | ...I can indeed create many constraint polyhedra - the trick though will be trimming them. |
| 20:29 | < Shoukanjuu> | I see |
| 20:29 | < Shoukanjuu> | 'Well, it was the way the conversation was set up :P |
| 20:34 | < Vornicus-Latens> | okay. I can get as many as... 7 planes from a particular planet; most of the time though there will be fewer. On the other hand, there are a /lot/ of different ways the systems could be set up. |
| 20:38 | | Vornicus-Latens is now known as Vornicus |
| 20:44 | | Attilla_ [~The.Attil@92.9.59.ns-21252] has joined #code |
| 20:44 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Ping Timeout] |
| 20:56 | | Vornicus is now known as Finerty |
| 21:00 | | Shoukanjuu [~Shoukanju@Nightstar-20375.dhcp.embarqhsd.net] has quit [Quit: Life is a game.. The object is to survive, to imprint in history your legacy. Go forth and show them what you are, and why you, your memory, and your soul will not be so easily vanquished!] |
| 21:01 | | Attilla_ is now known as Attilla |
| 21:02 | | Shoukanjuu [~Shoukanju@Nightstar-20375.dhcp.embarqhsd.net] has joined #code |
| 21:55 | | * gnolam stabs IIS with a rusty spork. |
| 22:13 | | You're now known as TheWatcher[zZzZ] |
| 22:17 | < ErikMesoy> | Who's doing the Dark Tower references? |
| 22:29 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Ping Timeout] |
| 22:37 | | Attilla [~The.Attil@92.9.59.ns-21252] has joined #code |
| 22:37 | | mode/#code [+o Attilla] by ChanServ |
| 22:41 | | ErikMesoy is now known as ErikMesoy|sleep |
| 22:52 | <@gnolam> | Arghl. No, you don't need to send a confirmation mail for every single name server change, you nincompoops. |
| 23:34 | | Attilla [~The.Attil@92.9.59.ns-21252] has quit [Ping Timeout] |
| 23:36 | | AnnoDomini [AnnoDomini@Nightstar-29230.neoplus.adsl.tpnet.pl] has quit [Quit: Drive me closer, I want to hit them with my sword!] |
| 23:51 | | Attilla [~The.Attil@92.9.59.ns-21252] has joined #code |
| 23:51 | | mode/#code [+o Attilla] by ChanServ |
| --- Log closed Mon Jun 30 00:00:15 2008 |