code logs -> 2007 -> Sun, 04 Mar 2007< code.20070303.log - code.20070305.log >
--- Log opened Sun Mar 04 00:00:51 2007
00:10 Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code
00:10 mode/#code [+o Chalcedon] by ChanServ
00:14 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
00:14 Janus is now known as Jan[draw]
00:32 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has joined #code
00:44 Chalcedon is now known as ChalcyCleaning
02:34 gnolam [Lenin@Nightstar-13557.8.5.253.se.wasadata.net] has quit [Quit: Z?]
03:01 MahalBEDD is now known as Mahal
03:14 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: Gone]
03:17 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
03:23 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Connection reset by peer]
03:25 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
04:01 McMartin[WonderCon] is now known as McMartin
04:14 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has quit [Quit: This computer has gone to sleep]
04:17 CommanderFrog is now known as ToxicFrog
04:22 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has joined #code
04:29 Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code
04:29 mode/#code [+o Chalcy] by ChanServ
04:30 ChalcyCleaning [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout]
04:30 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout]
04:30 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
04:56 McMartin [~mcmartin@Nightstar-5796.dsl.pltn13.sbcglobal.net] has quit [Quit: Kernel upgrade]
05:00 MyCatVerbs is now known as MyCatSleeps
05:28 Forjeh [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
05:29
< Jan[draw]>
Strange...
05:29 Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout]
05:30 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout]
05:30 Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code
05:30 mode/#code [+o Chalcedon] by ChanServ
05:31
< Jan[draw]>
A drawing can look perfectly fine, then when it's fliped horizontally, it's like looking into East Burlen.
05:35
<@Raif>
Berlin?
05:36 * Jan[draw] doesn't speak German.
05:37
< Jan[draw]>
Try speaking in American, it's the only language I understand.
05:37
< Doctor_Nick>
hurf durf blurf slurf
05:38 * Jan[draw] gives Dr. Nick another beer.
05:38 * Doctor_Nick spills it on himself and slumps on the credenza
05:42 ReivClass [~Reiverta@Nightstar-23347.ubs-dsl.xnet.co.nz] has joined #Code
05:43 ReivClass is now known as Reiver
05:44 Reiver is now known as NSGuest-795
05:48 NSGuest-795 is now known as Reiver
05:57 Jan[draw] [~Cerulean@Nightstar-10302.columbus.res.rr.com] has quit [Quit: "Self-esteem is overrated anyway. Most of the work in this country is done by people who hate themselves."]
06:26 Mahal [~Mahal@Nightstar-4998.worldnet.co.nz] has quit [Quit: fuk off i got biskit]
06:38 Vornicus-Latens is now known as Vornicus
07:16 ErikMesoy|sleep is now known as ErikMesoy
07:27 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has quit [Quit: This computer has gone to sleep]
07:32 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has joined #code
08:07 Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: ]
08:37 Forjeh [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: Gone]
08:38 GeekSoldier_ [Rob@Nightstar-2916.pools.arcor-ip.net] has joined #code
08:39 GeekSoldier [Rob@Nightstar-3246.pools.arcor-ip.net] has quit [Ping Timeout]
08:51 Mahal [~Mahal@Nightstar-4998.worldnet.co.nz] has joined #Code
08:51 mode/#code [+o Mahal] by ChanServ
09:26 GeekSoldier_ is now known as GeekSoldier
09:44 You're now known as TheWatcher
10:08
<@jerith>
Afternoon all.
10:15 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has left #code [Leaving]
10:23 Reiver [~Reiverta@Nightstar-23347.ubs-dsl.xnet.co.nz] has quit [Ping Timeout]
10:31 Reivtop [~reiver@Nightstar-27222.ubs-dsl.xnet.co.nz] has joined #Code
10:31
< Reivtop>
Dear Murphy:
10:32
< Reivtop>
While I understand you and I have not seen eye to eye in the past, and that you have delighted in destroying my livelihoods in the process, I feel that destroying my main boxen's PSU was a bit much.
10:32
< Reivtop>
But did you have to kill my LCD screen too?
11:11 Reivtop is now known as Reiver
12:11 Reiver is now known as ReivZzz
13:30 gnolam [Lenin@Nightstar-13557.8.5.253.se.wasadata.net] has joined #Code
14:00 ErikMesoy is now known as ErikMesoy|chibicatboy
14:51 ErikMesoy|chibicatboy is now known as ErikMesoy
14:55 Mahal is now known as MahalBEDD
15:09 * Serah PatPats ReivZzz.
15:09
< MyCatSleeps>
That was gracious.
15:09 MyCatSleeps is now known as MyCatVerbs
15:11 * Vornicus rewrites his longest-path finder to not prune locations you can get to in a longer time, because it turns out that that's bad.
15:12
<@jerith>
Generating infinite loops?
15:12
<@jerith>
Or something?
15:12
<@Vornicus>
No
15:12
<@Vornicus>
It is possible for a shorter path to a particular point to be better because it doesn't block entrance to certain segments of the graph.
15:13
<@jerith>
Umm... Are you only allowed to traverse each node once or something?
15:13
<@Vornicus>
Yep.
15:15
<@Vornicus>
Also, recursion and results caching are your tiny gods.
15:15
<@jerith>
Indeed.
15:16
<@jerith>
Very tiny, or more just sort of smallish?
15:16
<@Vornicus>
Just sort of smallish.
15:16
<@jerith>
:-)
15:17
<@Vornicus>
A depth first search with caching of failed lines has reduced what would be exponential time to almost-linear time.
15:17
<@Vornicus>
(for a given target node, I have to check all the paths that the node would be on for a duplicate of that node)
15:17
<@jerith>
:-D
15:19 * MyCatVerbs blinks.
15:19
< MyCatVerbs>
I thought the optimal running times for longest-path finding algorithms were exponential. oO
15:21
<@Vornicus>
If you don't store anything, I imagine so. Right now I'm looking at something around quartic time; if I were willing to more aggressively cache, it would be more along the lines of cubic time.
15:22
< MyCatVerbs>
Vornicus: "optimal," is defined to mean, "no algorithm can ever have a worst-case better than this." >_>
15:22
<@Vornicus>
Heh
15:23
< ErikMesoy>
Of course, no matter how many times the programmers and mathematicians say that you have to use the worst case as your baseline, PHBs will still give orders to code a moderately good system that does something nasty when given the worst case, such as taking indefinite time.
15:24
< MyCatVerbs>
ErikMesoy: that calls for excruciating hacks.
15:24
<@Vornicus>
I'm looking at entirely batshit storage, though - nodes * edges storage for paths, and good caching would be nodes^2 plus another nodes^2 intermittent.
15:24
< MyCatVerbs>
ErikMesoy: such as (if it's possible to do so cheaply) detecting that you're looking at something resembling your algorithm's worst case and switching algorithm to cope with it. =)
15:25
< ErikMesoy>
Oh, that's smart.
15:25
<@Vornicus>
But if you do that, you end up with, uh, 2 * nodes * nodes * edges time.
15:25
< ErikMesoy>
:D
15:26
< MyCatVerbs>
Vornicus: you're trading off storage for time?
15:26
<@Vornicus>
MCV: very much so.
15:27
< MyCatVerbs>
Fun. ^^
15:29
<@Vornicus>
But, yeah, my previous method was 1. crap, and 2. slower than this.
15:30
<@Vornicus>
By about an order of magnitude.
15:30
< MyCatVerbs>
Eh, just so long as you don't hit your swap file.
15:30
<@Vornicus>
I'm about at 50MB for an 1100-ish node graph, halfway through.
15:31
<@Vornicus>
This is in Python, though, so it's got overhead.
15:31
< Doctor_Nick>
N^N
15:31
< MyCatVerbs>
Vornicus: niiiice.
15:32
< Doctor_Nick>
im not even sure how to do an N^N function
15:33
< MyCatVerbs>
Doctor_Nick: some variations on bogosort exceed N^N.
15:33
<@Vornicus>
...though it has occurred to me that I haven't figured out how to get an actual path from the final data...
15:34
<@Vornicus>
...oh, duh, real simple. Adapt my valid-path-checking function to return a list or None, and if it returns a list, append current to it.
15:36
< MyCatVerbs>
This is the handy thing about languages like Python, especially when they have good C integration.
15:37
<@Vornicus>
...this is some wacky data, though - it dropped from 1067 to 993 live nodes after 8 moves, and has stayed there ever since. I suspect it might have a loop of 993 or thereabouts in it.
15:37
< MyCatVerbs>
If you need to make the whole program really really fast, it's entirely feasible to rewrite the path finding algorithm in C and link it in - *after* having prototyped and tuned it for complexity while working in Python.
15:40
<@Vornicus>
Indeed.
15:41
<@Vornicus>
I will eventually adapt this for my Settlers project; right now, however, I am content to watch it grind away on 1100 nodes of evil graph goodness.
15:42
< Doctor_Nick>
im not even sure how to do an N^N function
15:42
< Doctor_Nick>
oops
15:42
<@Vornicus>
You said that.
15:42
< Doctor_Nick>
i said that already
15:43
<@Vornicus>
Well, visiting n-tuples of n possible objects would do it.
15:52
<@Vornicus>
(so counting from 0 to 9,999,999,999, given 10...)
15:52
< MyCatVerbs>
Vornicus: Settlers?
15:53
<@Vornicus>
I'm making a Settlers of Catan webgame.
15:53
<@Vornicus>
Because the ones that are out there suck.
15:53
< MyCatVerbs>
Aaaand I have no idea what that is. *googles!*
15:54
< GeekSoldier>
only one of the greatest board games of all time.
15:54
< GeekSoldier>
:)
15:54
< gnolam>
Vornicus: what do you need fancy path finding for in SoC?
15:54
<@Vornicus>
gnolam: to find the longest road.
15:54
< GeekSoldier>
longroad?
15:55
< ErikMesoy>
Hmm. If you have memory, it seems like it might be better to count the increasing length of roads as they go, checking whenever a road is played.
15:55
< gnolam>
Since the number of roads is so limited, you shouldn't need anything special...
15:55
< gnolam>
IMO.
15:55
< GeekSoldier>
but roads may branch, so you'd need to check that too.
15:56
<@Vornicus>
Branch, and loop.
15:56 * GeekSoldier nods.
15:56
<@Vornicus>
My brother once won a game by creating a loop where once there was none.
15:56
<@Vornicus>
too bad, too, because had he not done so, I would have won by playing my third soldier.
16:00
<@Vornicus>
And with other considerations - Ships can move (and C&K's Diplomat card can remove or move roads), players can break your roads by playing settlements - it quickly becomes difficult to determine how these things affect road length directly. So much so that just performing an exhaustive search is simpler.
16:03
< gnolam>
But still, isn't it still just a DFS with a directed graph?
16:03
< gnolam>
-still, or something.
16:03
< gnolam>
Need. More. Caffeine.
16:04
< gnolam>
NEED.
16:04
<@Vornicus>
Yes. And that's what I'm doing.
16:04
<@Vornicus>
Well, except that I'm doing breadth-first, because it's cheaper to do all possible starting points at once.
16:08
<@Vornicus>
depth 650, 23000 cache items, 200,000 cache hits.
16:26 MyCatVerbs is now known as MyCatHungers
17:01 GeekSoldier is now known as Rosie
17:10 You're now known as TheWatcher[afk]
17:26 Rosie is now known as GeekSoldier
17:35 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
17:54 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: Gone]
17:56 Chalcedon [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code
17:56 mode/#code [+o Chalcedon] by ChanServ
18:32 You're now known as TheWatcher
18:38 MyCatHungers is now known as MyCatVerbs
18:56 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has joined #code
19:01 Forj [~Forj@Nightstar-869.bitstream.orcon.net.nz] has quit [Quit: Gone]
19:14 Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has joined #Code
19:18 AnnoDomini is now known as Kiers
19:22 Kiers is now known as KiersBrown
19:30
< GeekSoldier>
http://c2.com/cgi/wiki?LanguagesByKeyboard
19:47 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has joined #code
19:54 Syloq [Syloq@NetAdmin.Nightstar.Net] has joined #code
20:10 Doctor_Nick [~fdsaf@Nightstar-1992.9-67.se.res.rr.com] has quit [Quit: ]
20:11 EvilDarkLord is now known as Thokk
20:31 Syloq is now known as Syloqs-AFH
20:49 Syloqs-AFH [Syloq@NetAdmin.Nightstar.Net] has quit [Ping Timeout]
20:56 introc [~introc@Nightstar-24886.est.estpak.ee] has joined #Code
20:57
< introc>
offering safe and simple way to earn money! see the second half of the presentation: www.website.ws/whiteride/show
20:57 introc [~introc@Nightstar-24886.est.estpak.ee] has left #Code []
20:57
< GeekSoldier>
damn estonian pyramid schemes.
21:16
<@Vornicus>
On further analysis, the aggressive caching scheme I mentioned earlier doesn't really work - the last part of the algorithm is to actually generate a working path, and for that you need to traverse the entire graph - and for /that/ you'd better be caching the dead sections.
21:25 GeekSoldier [Rob@Nightstar-2916.pools.arcor-ip.net] has left #code []
21:34 Doctor_Nick [~fdsaf@Nightstar-1992.9-67.se.res.rr.com] has joined #code
21:35
< Doctor_Nick>
would any of you guys happen to have a copy of the 16-bit version of Turbo Debugger?
21:36 AnnoDomini [~farkoff@Nightstar-29071.neoplus.adsl.tpnet.pl] has joined #Code
21:36 AnnoDomini is now known as Kiers
21:37 KiersBrown [~farkoff@Nightstar-29017.neoplus.adsl.tpnet.pl] has quit [Ping Timeout]
21:51 ErikMesoy is now known as ErikMesoy|sleep
22:11 MyCatVerbs [~mycatownz@Nightstar-379.dsl.in-addr.zen.co.uk] has quit [Ping Timeout]
22:12 MyCatVerbs [~mycatownz@Nightstar-379.dsl.in-addr.zen.co.uk] has joined #code
22:28 Thaqui [~Thaqui@Nightstar-25354.jetstream.xtra.co.nz] has left #code [Leaving]
22:30 Thokk is now known as EvilDarkLord
22:41 Kiers is now known as AnnoDomini
22:55 Janus [~Cerulean@Nightstar-10302.columbus.res.rr.com] has quit [Quit: Jouets de Dieu, jouets de jouets, les jouets de me, na?tre Clair enfant voire.]
22:55 You're now known as TheWatcher[T-2]
22:58 You're now known as TheWatcher[zZzZ]
23:00 MahalBEDD is now known as Mahal
23:09 AnnoDomini [~farkoff@Nightstar-29071.neoplus.adsl.tpnet.pl] has quit [Quit: Those who slay dragons become dragons themselves. Stare into the abyss long enough, and it will stare back at you.]
23:23 ReivZzz is now known as Reiver
23:25 Chalcedon is now known as ChalcyOut
23:32 Chalcy [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has joined #code
23:32 mode/#code [+o Chalcy] by ChanServ
23:33 ChalcyOut [~Chalceon@Nightstar-869.bitstream.orcon.net.nz] has quit [Ping Timeout]
--- Log closed Mon Mar 05 00:00:51 2007
code logs -> 2007 -> Sun, 04 Mar 2007< code.20070303.log - code.20070305.log >