| --- Log opened Mon Mar 17 00:00:25 2008 |
| 00:18 | <@MyCatVerbs> | Why's MI considered so spidery anyway? |
| 00:19 | | * MyCatVerbs has never tried it, so has no reference for why everyone seems to consider it evil. |
| 00:20 | < Vornicus> | multiple inheritence is bad because it makes for precedence issues. |
| 00:20 | < Vornicus> | Especially when inheriting from a common ancestor, you will often end up with two definitions of a single method, and no particular agreement on which you use. |
| 00:22 | <@MyCatVerbs> | Dahhhhh. |
| 00:24 | <@MyCatVerbs> | Handy to know, thanks. |
| 00:35 | | gnolam [lenin@85.8.5.ns-20483] has quit [Quit: Z?] |
| 00:55 | <@McMartin> | Also fun is this situation: |
| 00:55 | <@McMartin> | class A { field f; } |
| 00:55 | <@McMartin> | class B extends A { ... } |
| 00:55 | <@McMartin> | class C extends A { ... } |
| 00:56 | <@McMartin> | class D extends B, C { ... } |
| 00:56 | <@McMartin> | How many copies of f does D have? |
| 00:58 | <@McMartin> | And more importantly, if you cast D to a C, and use mutators that mess with f, will they be reflected when you cast it to B and use *those* methods? |
| 00:58 | <@McMartin> | In C++ the answer is "by default, no, not even if you cast." |
| 00:58 | <@McMartin> | Er. Not even if you *don't* cast. |
| 00:59 | <@McMartin> | Unless you do virtual multiple inheritance, in which case they do. |
| 00:59 | < C_tiger> | TWWEBTLWN |
| 00:59 | <@McMartin> | whut |
| 00:59 | < C_tiger> | "Those words were English but that language was not" |
| 00:59 | <@McMartin> | Indeed it wasn't. The language was C++. |
| 01:00 | <@McMartin> | There are many reasons to hate it. The fact that it attempts to do all forms of MI, and that all of them are wrong, is one of those reasons. |
| 01:24 | | ReivUni [~82d94c4d@Nightstar-29731.dsl.in-addr.zen.co.uk] has joined #Code |
| 01:24 | < ReivUni> | Wheeeee |
| 01:26 | < C_tiger> | Wait til the drugs wear off. |
| 01:27 | < ReivUni> | Heh. I had two tests in one day for my compsci papers. They -clashed-, until about an hour ago when one of my lecturers let me do their test an hour early. |
| 01:27 | < ReivUni> | (I still have four hours of testing straight, but meh) |
| 01:35 | <@MyCatVerbs> | McMartin: so if I understand this correctly, the reason multiple-interfaces isn't a problem even though multiple-inheritances is is that interfaces don't specify behavoir (just signatures) and also never specify fields? |
| 01:36 | <@McMartin> | MCV: Quite so. |
| 01:37 | <@MyCatVerbs> | Even though you can effectively specify a field by adding getFoo+setFoo to an interface, and making the usual conditions part of the interface's contract? |
| 01:38 | <@McMartin> | Sure. But note then that you have your answer as to "how many versions are there" - to wit, always one. |
| 01:38 | <@MyCatVerbs> | Spiffy. |
| 01:39 | | * MyCatVerbs vaguely wonders if a straightforward ripoff of Haskell's typeclasses would be useful for a language like Java. |
| 01:41 | <@MyCatVerbs> | Since their capabilities are (AFAIK) a strict superset of Java's interfaces, but they have some really really nice additional features, too. Like you can define things like: if X is a member of typeclass Foo, Lists-of-X are members of typeclass Bar. |
| 01:43 | <@MyCatVerbs> | (For instance, printing of data structures is implemented like this. instance (Show a) => Show [a] where..., instance (Show a) => Show (Map a) where..., etc) |
| 01:54 | | * Vornicus makes King Hairy iteration. |
| 01:55 | | * Vornicus ...actually decides recursion is actually better. |
| 01:55 | < Vornicus> | otherwise I have to store iteration data in the structure, which isn't really that cool. |
| 02:10 | < Vornicus> | ...hell yes, it worked. |
| 02:10 | < Vornicus> | And shockingly quickly, too. |
| 02:11 | | * Vornicus thinks it worked, anyway. |
| 02:14 | <@ToxicFrog> | King Hairy? |
| 02:43 | < Vornicus> | I wanted to, for each square in a 9x8 grid, find a carcassonne tile that fits, and then move on to the next one, and so forth, until I run out of valid tiles. And then step backwards, find the next tile that fits, etc, etc. |
| 02:43 | < Vornicus> | THis would have required storing where i was in the iteration in each tile. |
| 02:46 | < ReivUni> | carcassonne tile? |
| 02:46 | | mode/#code [+o Vornicus] by Vornicus |
| 02:46 | <@Vornicus> | I switched it to recursive. http://vorn.dyndns.org/~vorn/carcassonne_random.py <--- messy, but it works. Decomment lines 76 and 86 to get a progress graph; said progress graph could stand some improvement. Anyway it's nondeterministic, and outputs a list of tile IDs and directions; the id tells you what tile to use, the direction tells you which way to point the canonical 'top' of the tile. |
| 02:47 | <@Vornicus> | http://vorn.dyndns.org/~vorn/carcassonne_square.txt <--- the formatted output. In a moment I'll be taking a picture. |
| 02:53 | <@Vornicus> | Carcassonne being a board game. |
| 02:53 | <@Vornicus> | well, a tile game. |
| 02:54 | <@Vornicus> | You build a city-state from tiles, and you must match the edges of the tiles with each other. |
| 02:56 | | ReivUni [~82d94c4d@Nightstar-29731.dsl.in-addr.zen.co.uk] has quit [Quit: CGI:IRC (EOF)] |
| 03:00 | <@Vornicus> | bah |
| 03:08 | | * Vornicus eyes his crappy-ass photography. The target isn't straight, and the lighting is awful. |
| 03:10 | <@Vornicus> | Granted, I did this without a flash and holding the camera pointing down over a table at arm's length to keep from blocking the existing light, at night, in the most light-absorbing room in the house. |
| 03:13 | <@Vornicus> | http://vorn.dyndns.org/~vorn/carcassonne_square.jpg |
| 03:17 | < JeffL> | Are those football-shaped cities allowed? |
| 03:17 | <@Vornicus> | Yes. |
| 03:17 | <@ToxicFrog> | Yes, but they only score half as much as larger ones. |
| 03:17 | <@Vornicus> | They're not worth as much, though |
| 03:17 | <@Vornicus> | Usually if you finish a city you get two points per tile and two points per shield |
| 03:17 | <@ToxicFrog> | Half as much per tile, that is |
| 03:18 | < JeffL> | It's scored? |
| 03:18 | <@Vornicus> | Yeah |
| 03:18 | < JeffL> | Does your algorithm find 1. A valid map or 2. The highest-scoring valid map? |
| 03:18 | <@Vornicus> | (the two croissant shaped cities in the west are worth 8 and 6 points, from north to south, because the top one has a shield) |
| 03:18 | <@Vornicus> | A valid map, that fits compactly into a rectangle. |
| 03:31 | | Derakon [~Derakon@Nightstar-6759.hsd1.mn.comcast.net] has joined #code |
| 03:31 | | mode/#code [+o Derakon] by ChanServ |
| 03:40 | <@Vornicus> | hey, look, the river tiles give me an odd number of city edges. Which means that I /can/ make it self-contained. |
| 03:42 | <@Derakon> | Rock. |
| 04:25 | | Vornicus [~vorn@ServicesOp.Nightstar.Net] has quit [Ping Timeout] |
| 04:30 | | Vornicus [~vorn@76.254.192.ns-13365] has joined #code |
| 04:30 | | mode/#code [+o Vornicus] by ChanServ |
| 04:32 | | Vornicus [~vorn@Admin.Nightstar.Net] has quit [Client exited] |
| 04:33 | | Vornicus [~vorn@76.254.192.ns-13365] has joined #code |
| 04:33 | | mode/#code [+o Vornicus] by ChanServ |
| 04:36 | | Vornicus [~vorn@Admin.Nightstar.Net] has quit [Client exited] |
| 04:37 | | Vornicus [~vorn@76.254.192.ns-13365] has joined #code |
| 04:37 | | mode/#code [+o Vornicus] by ChanServ |
| 04:40 | | Vornotron [~vorn@76.254.192.ns-13365] has joined #code |
| 04:43 | | Vornicus [~vorn@76.254.192.ns-13365] has quit [Quit: Leaving] |
| 05:15 | | Reltzik [Reltzik@Nightstar-6331.dsl.pltn13.sbcglobal.net] has joined #code |
| 05:24 | | Vornotron is now known as Vornicus |
| 05:26 | | * Vornicus randomly considers building his own Carcassonne art library. Would not be nearly as involved as Catan's. |
| 05:27 | < Reltzik> | You have a Catan art library? |
| 05:30 | < Vornicus> | Yes. |
| 05:30 | < Reltzik> | ... WHY? |
| 05:30 | < Vornicus> | For my planned Catan webgame |
| 05:30 | <@jerith> | Because he's writing a web-based Catan. |
| 05:31 | | mode/#code [+o Vornicus] by Vornicus |
| 05:31 | <@Vornicus> | http://unspeakablevorn.livejournal.com/30173.html#cutid1 |
| 05:31 | < Reltzik> | .... cool! |
| 05:48 | | AnnoDomini [AnnoDomini@83.21.54.ns-26803] has joined #Code |
| 05:48 | | mode/#code [+o AnnoDomini] by ChanServ |
| 05:55 | | Derakon is now known as Derakon[AFK] |
| 06:15 | | Reltzik [Reltzik@Nightstar-6331.dsl.pltn13.sbcglobal.net] has quit [Quit: DEATH TO THE UNDEAD!] |
| 06:31 | <@McMartin> | Sweet. |
| 06:31 | <@McMartin> | nfrotz works out of the box in Terminal.app |
| 06:34 | <@Vornicus> | Woohoo |
| 06:43 | <@McMartin> | Though there are some bizarre savegame-related errors I cannot reproduce, and so I am sending off debug versions to petitioners. |
| 06:57 | | Vornicus is now known as Vornicus-Latens |
| 07:04 | | GeekSoldier|bed is now known as GeekSoldier |
| 07:41 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code |
| 07:42 | | mode/#code [+o gnolam] by ChanServ |
| 08:11 | | GeekSoldier [~Rob@Nightstar-9482.dip.t-dialin.net] has quit [Ping Timeout] |
| 08:15 | | GeekSoldier [~Rob@91.18.96.ns-26390] has joined #code |
| 10:28 | | Thaqui [~Thaqui@Nightstar-123.jetstream.xtra.co.nz] has left #code [Leaving] |
| 11:14 | <@Reiver> | Oh, FWIW, Vorn: Please don't feel like you have to stop explaining things just because I quit ReivUni suddenly. |
| 11:15 | <@Reiver> | I come home and read it anyway, so. ;) |
| 11:35 | | AFKSkull [~none@Nightstar-7066.dyn.optonline.net] has joined #code |
| 11:36 | < AFKSkull> | Good morning. Anyone awake? |
| 11:39 | < GeekSoldier> | yes. |
| 11:47 | <@Reiver> | sup? |
| 11:52 | < AFKSkull> | Well, I ws wondring what the best approach to a simple(I think) project I have would be |
| 11:54 | < AFKSkull> | I wnt to create a basic importer between two formats of 3d scene. the base file is an extremely bare bones set of coordinates, and the destination format is a very similar ASCII format that has similar layout for coordinates but a lot more miscellenious information. |
| 11:54 | < AFKSkull> | Basically, I just want to take the variables from one and plug them into the right place on the other. |
| 11:54 | < GeekSoldier> | where does the miscellaneous information come from? |
| 11:55 | < AFKSkull> | well, in this case that information is irrelevant- I'll be using default information for it. |
| 11:56 | < AFKSkull> | I could probably do it in actionscript, but I recall vorn helping me with a better way to do something similar long ago. |
| 12:03 | < AFKSkull> | On the pastebin, 9w2Fx744 is the source format, with each item on it's own line. |
| 12:03 | < AFKSkull> | SDpku339 is the detination format |
| 12:03 | < AFKSkull> | *destination |
| 12:19 | | AbuDhabi [AnnoDomini@83.21.80.ns-27252] has joined #Code |
| 12:19 | | AnnoDomini [AnnoDomini@83.21.54.ns-26803] has quit [Ping Timeout] |
| 12:33 | <@Reiver> | hmm |
| 12:33 | <@Reiver> | So you basically want a string manipulation, ASCII? |
| 12:33 | < AFKSkull> | yup |
| 12:34 | <@Reiver> | hmn. |
| 12:34 | <@Reiver> | Well, the first and most obvious step is to identify all the different lines in the source and destination files (One assumes they are fundamentally the same when it comes to the Bits With Numbers Attached) |
| 12:35 | <@Reiver> | (If not, you may be in for a much larger world of hurt if it's not convinient newlines, but) |
| 12:38 | < AFKSkull> | well, it's fairly straightforward |
| 12:39 | < AFKSkull> | since in the file spat out of the source, every new line is the appropriate data |
| 12:39 | < AFKSkull> | all I need to do is format it suitibly and add the extra information, which will be repeated for each one |
| 12:45 | <@Reiver> | Aha |
| 12:45 | <@Reiver> | That is indeed very simple. |
| 12:46 | <@Reiver> | Only question really, is what languages are you proficient in, and do these lines have easily identifiable variables? |
| 12:49 | < AFKSkull> | I am particularly proficient in none |
| 12:50 | < AFKSkull> | and except for the fact that they are always in the same place(with dividers) then no |
| 12:50 | < AFKSkull> | I imagine the dividers would be plenty though |
| 12:50 | < AFKSkull> | I'll be using actionscript if there's nothing more suited to it |
| 12:51 | < AFKSkull> | which is basically javascript when it comes to text parsing, at least in 2.0 |
| 13:22 | | * ToxicFrog checks the format |
| 13:22 | <@ToxicFrog> | Yes, the dividers are plenty. |
| 13:23 | <@ToxicFrog> | Are there any constraints on what languages can be used? |
| 13:28 | < AFKSkull> | none whatsoever |
| 13:29 | <@ToxicFrog> | http://rafb.net/p/evV4Rw27.html here's a converter in awk |
| 13:30 | < AFKSkull> | with the caveat that I'll also need a link to a compiler |
| 13:33 | <@ToxicFrog> | Awk is an interpreted language. The interpreter comes standard on any *nix system. |
| 13:34 | <@ToxicFrog> | If you're on windows, mingw and cygwin both come with it, and gnuwin32 has a port of it as well |
| 13:34 | | * AFKSkull nods, is looking it up |
| 13:34 | <@ToxicFrog> | It will work best wrapped in bash so that you can easily go, say, for i in *.mdl; do awk -f convert.awk < "$i" > "${i/.mdl/.lwo}"; done |
| 13:35 | <@ToxicFrog> | And thus mass-convert your files. |
| 13:35 | | * ToxicFrog also has a converter in Lua, now, and one in bash itself isn't hard either |
| 13:36 | < AFKSkull> | well, what would be simplest, with the least installation and setup of stuff? |
| 13:36 | <@ToxicFrog> | In pretty much any language it's "read the input; split it apart on ';'; output the template with some pieces replaced by the stuff that you read;" |
| 13:37 | < AFKSkull> | I meant the implementation, not the script |
| 13:37 | <@ToxicFrog> | The script is the implementation. |
| 13:38 | <@ToxicFrog> | But I wasn't answering that, I was just commenting. |
| 13:38 | <@ToxicFrog> | I would say bash, except that it doesn't seem to be possible to get a windows port of bash outside of cygwin or msys |
| 13:38 | <@ToxicFrog> | Probably awk, then, if you're ok with invoking it from the dos shell |
| 13:39 | < AFKSkull> | I hve no particular problem with it |
| 13:39 | <@ToxicFrog> | http://gnuwin32.sourceforge.net/packages/gawk.htm |
| 13:40 | <@ToxicFrog> | The thing to keep in mind is that this script converts one file at a time; it's up to the shell to run it on all your model files. |
| 13:40 | <@ToxicFrog> | I think cmd.exe has something for that, but don't quote me on that. |
| 13:40 | < AFKSkull> | that's not a problem |
| 13:40 | < AFKSkull> | this isnt a model file, it's a scene file |
| 13:41 | < AFKSkull> | for translation, rotation, and scaling of a gigantic list of preconverted objects |
| 13:41 | <@ToxicFrog> | Aah |
| 13:41 | <@ToxicFrog> | Oh, it's a list? |
| 13:41 | <@ToxicFrog> | It |
| 13:41 | < AFKSkull> | yup |
| 13:41 | <@ToxicFrog> | s not one object per file? |
| 13:41 | <@ToxicFrog> | Ok, let me make some quick modifications to the script then~ |
| 13:41 | < AFKSkull> | no, it's several thousand objects |
| 13:41 | < AFKSkull> | one per line |
| 13:41 | <@ToxicFrog> | Does each one need to go into a different file? |
| 13:42 | < AFKSkull> | no |
| 13:42 | < AFKSkull> | all into one |
| 13:42 | <@ToxicFrog> | Awesome. |
| 13:42 | <@ToxicFrog> | In that case, just remove the "exit" from the end of the awk script |
| 13:42 | < AFKSkull> | basically, just expanding the list, which is basically a stripped down scene file, into the full thing |
| 13:43 | <@ToxicFrog> | And then you run it using: awk.exe -f <script file> <scene file> > <output file> |
| 13:43 | < AFKSkull> | the numbers don't translate properly, but I already write a script internally to lightwave to scale and move it once it's in the right format |
| 13:43 | <@ToxicFrog> | Eg, awk -f convert.awk scene_stripped scene.lws |
| 13:43 | <@ToxicFrog> | Er |
| 13:43 | <@ToxicFrog> | Eg, awk -f convert.awk scene_stripped > scene.lws |
| 13:43 | < AFKSkull> | gotcha |
| 13:43 | <@ToxicFrog> | Forgot the redirect operator ?? |
| 13:44 | <@ToxicFrog> | Without the > scene.lws it will just dump it all to the terminal for you to take a look. |
| 13:44 | < AFKSkull> | and then I just get to figure out which of the thousands of objects I want to keep @_@ |
| 13:44 | <@ToxicFrog> | I note that getting it to adjust the coordinates and suchlike at the same time wouldn't be hard, either. |
| 13:44 | < AFKSkull> | well, it's not always the same |
| 13:44 | <@ToxicFrog> | Aah |
| 13:44 | <@ToxicFrog> | Probably more trouble than it's worth, then |
| 13:45 | <@ToxicFrog> | Since you already have a script for that. |
| 13:45 | < AFKSkull> | so what I did was make it so I'd adjust a single reference object, and then it would automatically propagate the changes |
| 13:46 | < AFKSkull> | but it only works if I have them loaded properly and placed relative to their initial relative spots |
| 13:47 | | * AFKSkull gives it a try |
| 13:48 | | mode/#code [+vvvvv AbuDhabi AFKSkull Attilla C_tiger GeekSoldier] by ToxicFrog |
| 13:48 | | mode/#code [+iRvvv jerith Kazriko Pi] by ToxicFrog |
| 13:49 | | mode/#code [+vvv JeffL Raif Serah] by ToxicFrog |
| 13:49 | | mode/#code [-iR] by ToxicFrog |
| 13:58 | <@ToxicFrog> | Well, I'm off to class. Later! |
| 14:01 | <+AFKSkull> | ok, it works, but doesn't actually save the file |
| 14:01 | <+AFKSkull> | thanks a ton though, Im sure I'll figure it out now^_^ |
| 15:27 | <@ToxicFrog> | The > filename is meant to save it |
| 15:27 | <@ToxicFrog> | It redirects output to that file |
| 15:34 | <+AFKSkull> | ah, right |
| 15:34 | <+AFKSkull> | I got it now, thanks |
| 15:37 | <+AFKSkull> | aha, seems to be working perfectly |
| 15:41 | <+AFKSkull> | albeit with absolutely insane results, but that will have to be worked out in the model exporter |
| 15:42 | <+AFKSkull> | thanks a lot^_^ |
| 15:43 | <@ToxicFrog> | No problem |
| 15:44 | <@ToxicFrog> | It really is a two minute job in most languages. |
| 15:45 | <+AFKSkull> | to be fair, the hardest part of doing it in actionscript would be getting flash to load the file |
| 15:46 | <+AFKSkull> | but you saved me a ton of work, definately |
| 15:58 | <+AFKSkull> | http://pinkhair3d.blogspot.com/2008/03/wee.html |
| 17:32 | | mode/#code [+oooooo AbuDhabi AFKSkull Attilla C_tiger GeekSoldier JeffL] by Vornicus-Latens |
| 17:32 | | mode/#code [+oooo Kazriko Pi Raif Serah] by Vornicus-Latens |
| 17:38 | | * AbuDhabi crackles with power. |
| 17:52 | <@C_tiger> | Ooooh. |
| 17:56 | | Vornicus-Latens is now known as Vornicus |
| 17:58 | | * Vornicus updates the access list. |
| 18:12 | <@C_tiger> | Does this channel get spammed a lot such that it needs to be +U? |
| 18:13 | <@C_tiger> | Or is that just tradition? |
| 18:14 | <@AbuDhabi> | Tradition. We don't get very many spammers on this network. Or, rather, they all hang out in #oubliette. |
| 18:14 | <@C_tiger> | Ok, just curious. |
| 18:15 | <@C_tiger> | I mean I don't get spammers in the other channels I'm in but I thought maybe the more "generic" channels may be different. |
| 18:16 | | * Vornicus isn't exactly sure why it's +U any more. |
| 18:19 | <@GeekSoldier> | there had been a small problem in the past with drive-through spam. |
| 18:20 | <@Vornicus> | ah, so |
| 18:21 | <@Vornicus> | This is one of the largest public channels. |
| 18:29 | <@C_tiger> | Wow, apparently #fleet and #crfh and #code are among the top 5 non-secret channels. |
| 18:30 | <@AbuDhabi> | #fleet and #crfh are secret. |
| 18:30 | <@C_tiger> | Oh, right, I'm in them, that's why they show up in /list |
| 18:48 | <@Vornicus> | um |
| 18:49 | <@Vornicus> | ..heh |
| 18:49 | <@Vornicus> | Anno got there first |
| 19:19 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout] |
| 19:19 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code |
| 19:19 | | mode/#code [+o gnolam] by ChanServ |
| 19:26 | | * Vornicus fiddles with the carcassonne thingy, changing the outputter and the match checker. |
| 19:27 | | * Vornicus also adds cloister data. |
| 19:38 | | AbuDhabi is now known as Steven |
| 19:46 | <@jerith> | Is #crfh secret? |
| 19:47 | <@jerith> | I'd've thought a webcomic channel would be public. |
| 19:52 | | * Vornicus rejiggers his progress output, watches it tick over, is so glad he knows that backspace trick. |
| 19:55 | <@jerith> | Backspace trick? |
| 19:56 | <@Vornicus> | To make a progress bar. |
| 19:56 | <@jerith> | Ah. |
| 19:56 | <@Vornicus> | If you want to rewrite something, you backspace over it, flush the backspaces, and then write what you want. |
| 19:56 | | * jerith nods. |
| 19:58 | <@Vornicus> | So I've got my carcassonne progress thingy buzzing through possibilities, and telling me what fraction of the possibilities it's gone through for each individual tile. |
| 20:03 | | GeekSoldier is now known as GeekSoldier|bed |
| 20:15 | <@Vornicus> | THis one's taking a while, because it's got other things to deal with. |
| 20:16 | <@Vornicus> | Other restrictions, that is. The original did not take nearly so long, because it didn't have to restrict the search space so the river doesn't go off the edge. |
| 20:46 | <@jerith> | http://www.classnamer.com/ |
| 20:48 | <@Vornicus> | RepublicanHashSubscriber. |
| 20:48 | <@Vornicus> | Yeah... |
| 20:48 | <@gnolam> | Hah. |
| 20:49 | | * gnolam is so going to use that in his next assignment. |
| 20:53 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout] |
| 20:55 | | gnolam [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code |
| 20:55 | | mode/#code [+o gnolam] by ChanServ |
| 21:05 | | Jeff [Kaline@Nightstar-12251.dsl.sndg02.pacbell.net] has joined #code |
| 21:07 | | JeffL [JPL@Nightstar-12251.dsl.sndg02.pacbell.net] has quit [Ping Timeout] |
| 21:16 | <@Vornicus> | gah. This is taking an age. |
| 21:19 | <@Vornicus> | ...I imagine that this would go faster if I could path the river first. |
| 21:27 | | Reiver [~reaverta@Admin.Nightstar.Net] has quit [Ping Timeout] |
| 21:29 | | Reiver [~reaverta@Nightstar-8829.xdsl.xnet.co.nz] has joined #Code |
| 21:29 | | mode/#code [+o Reiver] by ChanServ |
| 21:30 | <@gnolam> | What are you doing, exactly? Pre-placing a Carcassone board? |
| 21:30 | <@Vornicus> | Trying to build a Carcassonne board that fits into a rectangle |
| 21:32 | <@gnolam> | Ah. |
| 21:32 | <@gnolam> | Just to see if it can be done? :) |
| 21:32 | <@Vornicus> | I'm pretty certain it can be done. It's just factorial complexity. |
| 21:35 | | Steven is now known as AnnoDomini |
| 21:37 | <@Vornicus> | Right now I'm watching one buzz at around 80/84 complete, and trying to figure out how to prebuild the river. |
| 21:42 | <@Vornicus> | Because it's pretty clear that what's happening is I'm getting rivers that can't hope to match up, and so in order to fix it it has to go all the way back to the start of the river and try again. |
| 21:43 | <@Vornicus> | Which is an incredible amount of work because the algorithm doesn't understand that. |
| 21:43 | <@gnolam> | Ah. |
| 21:44 | <@Vornicus> | (the algorithm is just a recursive thing - it tries all the available tiles in the next available space, and if it finds one that works, it goes on to the next, and if it doesn't, it backs up and continues that one's iteration) |
| 21:44 | <@C_tiger> | Ick. |
| 21:45 | <@Vornicus> | Which works fine, when you don't have the heavily restricted river tiles in there; it gets an answer for river-free in a few seconds. |
| 21:45 | <@Vornicus> | --and that's in unoptimized python with a time-consuming debug. |
| 21:45 | <@C_tiger> | Can you prebuild river "sections"? |
| 21:46 | <@Vornicus> | I think what I'm going to do is try pathing and placing the river /first/, and then convincing the recursive section to walk over it. |
| 21:46 | <@C_tiger> | Ok, that works too. |
| 21:46 | <@Vornicus> | To step over it, rather, so it doesn't try placing anything on the river. |
| 21:49 | <@C_tiger> | Well, the river should divide the board right? so can you use a "fill" algorithm, treating the river as a boundary? |
| 21:50 | <@Vornicus> | The river doesn't really divide the board. |
| 21:50 | <@Vornicus> | The river is twelve tiles, with a beginning and an end. |
| 21:51 | <@C_tiger> | regardless. |
| 21:51 | <@C_tiger> | I mean by treating the river as a boundary, you avoid all the problems of disjointed tiles when you "walk over" it. |
| 21:52 | <@Vornicus> | Then comes the problem of "which order do I cover the tiles in" and "where exactly in the rectangle is the river" |
| 21:53 | <@C_tiger> | I thought you were prebuilding the river. |
| 21:53 | <@Vornicus> | Right |
| 21:53 | <@C_tiger> | So you know where the river is. |
| 21:53 | <@C_tiger> | I'm just suggesting "walk around" rather than "walk over" |
| 21:54 | <@Vornicus> | thing is, using the walk-around I need to use the river's position and spiral out from it (this is actually pretty hard) and then filter out things that aren't in the rectangle |
| 21:54 | <@C_tiger> | No, I mean still start in the corner. |
| 21:55 | <@C_tiger> | But when you hit the river, fill rather than jump. |
| 21:55 | | * Vornicus thinks |
| 21:55 | <@C_tiger> | The wall is a boundary too. |
| 21:55 | <@C_tiger> | It's just a recursive fill algorithm. |
| 21:55 | <@Vornicus> | ...I don't even know any fill algorithms. |
| 21:56 | <@C_tiger> | Isn't it just for each tile, fork for each tiles around it. |
| 21:56 | <@C_tiger> | *each tile |
| 21:56 | <@Vornicus> | ah, so |
| 21:56 | <@C_tiger> | If a tile is completely surrounded, stop. |
| 21:57 | <@Vornicus> | THing is, it's poosible for the river to touch two edges and this block off some of the tiles as regards one corner. |
| 21:58 | <@Vornicus> | --indeed, I can create a scenario where there are as many as three disjoint sections. |
| 21:58 | <@C_tiger> | But that's an easy case to check for. Fill from one corner. Test other corners as needed. |
| 21:58 | <@C_tiger> | The only issue is if you have a circular river. |
| 21:59 | <@Vornicus> | I think I'm joust going to do skip. It's non-branching recursive, then. |
| 22:00 | <@C_tiger> | Ok. |
| 22:00 | <@C_tiger> | *shrug* it was an idea. |
| 22:00 | <@Vornicus> | (and in this case, non-branching recursive is preferable because branches can block each other off...) |
| 22:04 | | Brother_Willibald [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has joined #Code |
| 22:04 | | gnolam is now known as NSGuest-5587 |
| 22:06 | | NSGuest-5587 [lenin@Nightstar-10613.8.5.253.static.se.wasadata.net] has quit [Ping Timeout] |
| 22:07 | | Brother_Willibald is now known as gnolam |
| 22:28 | | AnnoDomini [AnnoDomini@83.21.80.ns-27252] has quit [Quit: I. AM. BEOWULF.] |
| 23:35 | <@Vornicus> | ...alternatively, I can path the river, and then add checks to the match thing to make sure that river tiles get into the path. |
| 23:36 | | The_Mighty_END [~Zeratul@58.6.44.ns-20543] has joined #code |
| 23:47 | | Reiver [~reaverta@Nightstar-8829.xdsl.xnet.co.nz] has quit [Quit: Changing servers] |
| 23:48 | | Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code |
| 23:48 | | mode/#code [+o Reiver] by ChanServ |
| --- Log closed Tue Mar 18 00:00:35 2008 |