| --- Log opened Fri Dec 04 00:00:00 2009 |
| 00:02 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 00:07 | | Netsplit over, joins: Vornicus, simon`, @jerith, GeekSoldier, crem |
| 00:07 | | Netsplit over, joins: EvilDarkLord, SmithKurosaki, KazWork, Namegduf, MyCatVerbs, Alek |
| 00:09 | | You're now known as TheWatcher[T-2] |
| 00:10 | | Netsplit *.net <-> *.split quits: Alek, Namegduf, SmithKurosaki, EvilDarkLord, MyCatVerbs, KazWork |
| 00:10 | | * gnolam headdesks. |
| 00:11 | < gnolam> | One of my ugly hacks coming back to bite me in the ass. Who'da thunk it? |
| 00:11 | < Vornicus> | Gee. |
| 00:12 | | Netsplit over, joins: EvilDarkLord, SmithKurosaki, KazWork, Namegduf, MyCatVerbs, Alek |
| 00:13 | | You're now known as TheWatcher[zZzZ] |
| 00:18 | | Derakon[work] [Derakon@Nightstar-d44d635e.ucsf.edu] has quit [[NS] Quit: Leaving] |
| 00:27 | < gnolam> | (I'm currently working on underwater rendering. To render the water surface I just switch to culling front faces and reverse the normal when calculating the cubemap reflection. For some reason, when I got below a certain (very shallow!) depth, everything got grayed out. I spent about an hour going over every line of my shaders before I realized it was the giant "extend to horizon" quads I put in place because I haven't implemented an adaptive |
| 00:56 | | Derakon[AFK] [Derakon@Nightstar-5abd3ac9.ca.comcast.net] has joined #code |
| 00:57 | | Derakon[AFK] is now known as Derakon |
| 01:06 | < Vornicus> | Hokay so. Headers now fixed; now I need to get /chapter and /section to pull themselves out of columns. This might be pretty easy, actually. |
| 01:33 | | Attilla [The.Attilla@FBC920.398CA6.F5AD58.B7D1D3] has quit [Connection reset by peer] |
| 01:49 | < Vornicus> | Or not exactly. Got it, sort of, but now my body text is huge. |
| 01:50 | < gnolam> | That's just your distorted <body><img> talking. |
| 01:50 | | * gnolam hides. |
| 01:51 | | * Vornicus shoots gnomlam, points out that he's working in LaTeX |
| 01:51 | < Vornicus> | (also the font size is literally called \Huge, so) |
| 01:58 | < Vornicus> | There that goes. Okay. |
| 02:01 | < Vornicus> | Now that I actually have some idea of what LaTeX is actually capable of, these others will be relatively easy. |
| 02:02 | | Kazriko [kaz@Nightstar-e09690fa.client.bresnan.net] has joined #code |
| 02:29 | | Kazriko [kaz@Nightstar-e09690fa.client.bresnan.net] has quit [[NS] Quit: Leaving] |
| 02:47 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has quit [[NS] Quit: Z?] |
| 03:31 | < Derakon> | Holy blap, the Angband source is 137681 lines of C. |
| 04:36 | < Alek> | that actually sounds.... small. huh. |
| 04:37 | < Vornicus> | It's not very. |
| 04:40 | | * Vornicus fiddles with LaTeX, is pretty sure he actually sort of gets it now. |
| 04:51 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 04:53 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 04:57 | < Derakon> | Alek: for some perspective, Jetblade, counting all configuration modules, is just under 8k lines of Python. I expect that the finished game would be maybe two to three times that, tops. |
| 05:16 | | * Derakon figures out how to automatically make every item in the game be identified when he finds it. |
| 05:16 | < Derakon> | So that's the first step of the tweak I wanted...the rest'll be harder, of course. |
| 05:20 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has joined #code |
| 05:26 | < Derakon> | I note to my considerable dismay that the Angband source code uses tabs for indenting instead of spaces. |
| 06:32 | | Derakon is now known as Derakon[AFK] |
| 06:56 | < McMartin> | Wow. I guess having state really helps. |
| 06:56 | | * McMartin solves #126 in 90s, finds everyone else did basically exactly the same thing but solved it in several hundred ms instead. |
| 06:57 | < McMartin> | Including the Haskell code, which uses IOArray monads. Cheating! |
| 07:16 | < Vornicus> | Heh |
| 07:33 | | AnnoDomini [annodomini@Nightstar-6ee2c5d2.adsl.tpnet.pl] has joined #code |
| 07:54 | | * McMartin also blows out his RAM a few times before getting the lazy evaluation under control. |
| 08:28 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: ] |
| 08:30 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 08:39 | < Vornicus> | So, next target: make Problem, Subproblem, and Solution work together. |
| 08:59 | | Rhamphoryncus [rhamph@Nightstar-a62bd960.abhsia.telus.net] has quit [Client exited] |
| 08:59 | | You're now known as TheWatcher |
| 09:03 | < McMartin> | Hrm. There's gotta be a better way to do #127, but damned if I know what it is. |
| 09:03 | < McMartin> | The part where you actually factorize takes two seconds, and then dealing with the results seems to be taking upwards of half an hour. |
| 09:03 | | Vornicus is now known as Vornicus-Latens |
| 09:06 | < Vornicus-Latens> | rad(abc) = rad(a) * rad(b) * rad(c), if you manage to get that first bit done. |
| 09:08 | < McMartin> | Yes. |
| 09:08 | < McMartin> | Assuming that they're all relatively prime. Which they are. |
| 09:08 | < McMartin> | And since the rad(abc) < c check is cheaper than the relatively prime check, you do it first. |
| 09:08 | < Vornicus-Latens> | Similarly you only need to check two pairs of relative primeness - the third is implied. |
| 09:09 | < McMartin> | It's more that (ignoring a couple of lame prefilters) there appear to be 7.2M candidates. |
| 09:09 | < McMartin> | And they seem to get noticably denser once c > 10000 or so. |
| 09:10 | < McMartin> | If 2*rad(c) >= c we can also discard that c instantly. |
| 09:11 | < Vornicus-Latens> | --only happens if radc is squarefree or almost. |
| 09:12 | < McMartin> | Yeah, like I said, lame filter. |
| 09:12 | < Vornicus-Latens> | lame. |
| 09:12 | < McMartin> | But I see no guarantee that any larger factor won't possibly miss things. |
| 09:12 | < McMartin> | I don't see why you only have to check two pairs though, in the general case. 2, 3, 15 has ab and ac but not bc, and 2, 3, 8 has ab and bc but not ac. |
| 09:13 | < Vornicus-Latens> | Because you won't get either of those! |
| 09:13 | < Vornicus-Latens> | You'll get 2, 13, 15, or 3, 12, 15. |
| 09:14 | < McMartin> | While it is true that I won't get either of those, I have no a priori reason to believe that a property of sums will have any bearing on a property of division. |
| 09:14 | < Vornicus-Latens> | If a and c are relatively prime, then... oh, shit, I think you can do /only one/ gcd and have it work. |
| 09:14 | < McMartin> | Unless there is, in fact, a proof that if a and c are relatively prime then a and (c-a) are relatively prime and (c-a) and c are. |
| 09:14 | < Vornicus-Latens> | There is! |
| 09:15 | < Vornicus-Latens> | In fact it is embedded in Euclid's original algorithm - which was repeated subtractions, not division with remainder. |
| 09:16 | < McMartin> | Even so, that sounds like it's not going to be a particularly huge improvement. It would be, at best, a 75% speedup. |
| 09:16 | < McMartin> | That isn't nearly enough to make this finish in under a minute. |
| 09:16 | < McMartin> | I suspect the existence of a better filter. |
| 09:16 | < McMartin> | *pre-filter, rather. |
| 09:17 | < McMartin> | (Asymptotically, it's going from n^2 log n to n^2, which is real, but still kind of laughable.) |
| 09:17 | < McMartin> | (I'm guessing there's something that drops it to O(n) or below.) |
| 09:17 | < McMartin> | Also assuming I'm correctly remembering that gcd is O(log n). |
| 09:18 | < Vornicus-Latens> | If two numbers share a factor then you can, in any addition or subtraction involving those two numbers, factor that out - and you're left with integers, and the integers are closed, so the result (once you pull the factor back in) is integer. |
| 09:19 | < Vornicus-Latens> | er, also shares the factor. |
| 09:19 | < Vornicus-Latens> | You're right, it's not much improvement. Except... |
| 09:22 | | * McMartin 's dumb version is up to all values where c is below 83,000 now. |
| 09:22 | < Vornicus-Latens> | Hm, how would I actually end up doing this? c, having been factored, can be placed in a bunch of seive piles; you can then take a set of possible a values and, instead of gcding, just difference out the lower gcds. |
| 09:22 | < McMartin> | Seives are, of course, made of inconceivable sadness in Haskell. |
| 09:23 | < Vornicus-Latens> | :( |
| 09:23 | < McMartin> | I'm currently using a princely 13MB of memory for all my data structures, and they're all firmly write-once except for the actual output stream. |
| 09:24 | < McMartin> | The other thing is that performing the seive would end up being at least as if not more expensive than just doing the linear check. |
| 09:24 | < McMartin> | Because you'd be doing set subtraction. Which is O(n). |
| 09:25 | < McMartin> | Which is the same as a list filter. |
| 09:26 | | * Vornicus-Latens asks Wolfram Alpha. |
| 09:26 | < McMartin> | The trick is almost certainly in the choice of a, though, not c. I'm watching my results scroll by and it looks like perhaps one in three values of c actually has at least one abc-hit in it. |
| 09:26 | < McMartin> | And they have *lots*. |
| 09:28 | < McMartin> | c = 88625, for instance, has 17,465 corresponding abc hits, assuming my algo isn't buggy. |
| 09:28 | < Vornicus-Latens> | mcwhut |
| 09:29 | < McMartin> | (And it got the right answer for c<1000, so...) |
| 09:29 | | * McMartin is now past 90k. |
| 09:29 | | * McMartin is still only using 3% of his 512MB of RAM. |
| 09:32 | < McMartin> | I guess they're rarer than Pythagorean Triples, but. |
| 09:33 | < Vornicus-Latens> | Triples at least you can generate. |
| 09:34 | < McMartin> | Yeah, and it won't surprise me if you can generate these too. |
| 09:34 | < McMartin> | (96k) |
| 09:39 | | * Vornicus-Latens actually goes to bed. |
| 09:53 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has quit [Operation timed out] |
| 10:00 | | Attilla [The.Attilla@FBC920.398CA6.F5AD58.B7D1D3] has joined #code |
| 10:20 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has quit [Connection closed] |
| 10:20 | < AnnoDomini> | It seems that the new euphemism for failing a test is 'join the Jedi Academy' in my Uni. |
| 11:07 | | MyCatSchemes [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 11:28 | | gnolam [lenin@Nightstar-38637aa0.priv.bahnhof.se] has joined #code |
| 11:36 | | Tarinaky [Tarinaky@Nightstar-7b0c05ad.adsl.virginmedia.net] has quit [Operation timed out] |
| 11:41 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
| 11:50 | | Tarinaky [Tarinaky@Nightstar-cd385d3b.adsl.virginmedia.net] has joined #code |
| 12:05 | | AbuDhabi [annodomini@Nightstar-bbe88297.adsl.tpnet.pl] has joined #code |
| 12:07 | | AnnoDomini [annodomini@Nightstar-6ee2c5d2.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 12:25 | < simon`> | I wonder why they think Jedis fail tests. |
| 13:39 | < AbuDhabi> | Microsoft products are shit. |
| 13:39 | < AbuDhabi> | That is all. |
| 13:47 | < simon`> | there are six palindrome unixtimes from now and a week ahead |
| 14:06 | < gnolam> | Oh, evil Satan |
| 14:06 | < gnolam> | oscillate my metallic |
| 14:06 | < gnolam> | sonatas live, ho! |
| 14:21 | | Tarinaky [Tarinaky@Nightstar-cd385d3b.adsl.virginmedia.net] has quit [Ping timeout: 121 seconds] |
| 14:24 | | crem [moo@Nightstar-8ca3eea7.adsl.mgts.by] has quit [Client closed the connection] |
| 14:29 | | crem [moo@Nightstar-8ca3eea7.adsl.mgts.by] has joined #code |
| 14:35 | | AnnoDomini [annodomini@Nightstar-13a67611.adsl.tpnet.pl] has joined #code |
| 14:35 | | Tarinaky [Tarinaky@Nightstar-2b47a2d9.adsl.virginmedia.net] has joined #code |
| 14:37 | | AbuDhabi [annodomini@Nightstar-bbe88297.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 14:40 | | AbuDhabi [annodomini@Nightstar-1e44073d.adsl.tpnet.pl] has joined #code |
| 14:42 | | AnnoDomini [annodomini@Nightstar-13a67611.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 14:46 | | AnnoDomini [annodomini@Nightstar-a8ece2d3.adsl.tpnet.pl] has joined #code |
| 14:48 | | AbuDhabi [annodomini@Nightstar-1e44073d.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 14:50 | < gnolam> | Oh hey, Bing is actuall better than Google for searching for C++ stuff. The ++s appear to have actual meaning. |
| 14:50 | < gnolam> | +y |
| 14:51 | | AbuDhabi [annodomini@Nightstar-7f100388.adsl.tpnet.pl] has joined #code |
| 14:51 | | AnnoDomini [annodomini@Nightstar-a8ece2d3.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 15:14 | | Tarinaky [Tarinaky@Nightstar-2b47a2d9.adsl.virginmedia.net] has quit [Ping timeout: 121 seconds] |
| 15:20 | | AnnoDomini [annodomini@Nightstar-f75657b1.adsl.tpnet.pl] has joined #code |
| 15:22 | | AbuDhabi [annodomini@Nightstar-7f100388.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 15:25 | < Attilla> | Maybe it will also recognise other punctuation. |
| 15:28 | | Tarinaky [Tarinaky@Nightstar-41f555ed.adsl.virginmedia.net] has joined #code |
| 15:30 | | AbuDhabi [annodomini@Nightstar-b87577af.adsl.tpnet.pl] has joined #code |
| 15:32 | | AnnoDomini [annodomini@Nightstar-f75657b1.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 15:41 | | Tarinaky [Tarinaky@Nightstar-41f555ed.adsl.virginmedia.net] has quit [Ping timeout: 121 seconds] |
| 15:55 | | Tarinaky [Tarinaky@Nightstar-52356a03.adsl.virginmedia.net] has joined #code |
| 16:02 | < gnolam> | But the main thing is that it doesn't distract med with shitty DHTML. |
| 16:08 | | * gnolam stabs Google some more for putting the final nail in e-mail etiquette's coffin. |
| 16:14 | < Attilla> | At least it's not XHTML! :V |
| 16:14 | < Attilla> | Let's have a clock... THAT FOLLOWS YOUR POINTER |
| 16:15 | < simon`> | I hate email |
| 16:16 | < simon`> | gmail's web interface has made me forget how horrible email is, but now that I've tried getting a client up and running, I realize how dissatisfactory the clients I know are. |
| 16:22 | | * AbuDhabi uses the old HTML interface on Gmail. |
| 16:34 | < gnolam> | You're pretty much forced to, if you want it to run at any speed. :P |
| 16:39 | < gnolam> | Attilla: Err... what? |
| 16:46 | < Attilla> | Some random XHTML scripts I recall when I first started doing html in secondary school, this one site had tons of them |
| 16:47 | < Attilla> | Including an ASCII clock that followed your cursor around |
| 16:48 | < gnolam> | Err... XHTML is just HTML with stricter XML markup. |
| 16:50 | < Attilla> | fine, I meant DHTML? |
| 16:53 | < gnolam> | Yes. |
| 16:53 | < Attilla> | Huh, that's what you mentioned. Then fine :| |
| 16:53 | | * Attilla harumphs |
| 17:02 | < gnolam> | Javascript - the scource of the World Wide Web. |
| 17:02 | < Alek> | scourge |
| 17:03 | < gnolam> | ... yes, thank you. |
| 17:04 | < gnolam> | Need more caffeine. |
| 17:04 | < gnolam> | SPAWN MORE TEA |
| 17:05 | < AbuDhabi> | It's not like tea grows on trees. |
| 17:10 | < Alek> | oh wait... |
| 17:11 | < Alek> | ¬_¬ |
| 17:19 | | Rhamphoryncus [rhamph@Nightstar-a62bd960.abhsia.telus.net] has joined #code |
| 18:02 | < gnolam> | There. Finally got my meeting minutes transcribed and uploaded. |
| 18:02 | < gnolam> | Now: to seek compensation. |
| 18:05 | < gnolam> | Only I have no idea how to do so. Hmm. |
| 18:28 | | SmithKurosaki [Smith@Nightstar-e2478917.cable.rogers.com] has quit [Ping timeout: 121 seconds] |
| 18:28 | | SmithKurosaki [Smith@Nightstar-e2478917.cable.rogers.com] has joined #code |
| 18:45 | | AgentIchiro [Smith@Nightstar-f933638e.dsl.teksavvy.com] has joined #code |
| 18:45 | | SmithKurosaki [Smith@Nightstar-e2478917.cable.rogers.com] has quit [Ping timeout: 121 seconds] |
| 18:46 | | AgentIchiro is now known as SmithKurosaki |
| 18:49 | | MyCatSchemes [mycatverbs@Nightstar-3b2c2db2.bethere.co.uk] has quit [[NS] Quit: Leaving] |
| 20:10 | < McMartin> | Oy. I was getting wrong answers due to internal integer overflow. Silliness. |
| 20:18 | < simon`> | hehe |
| 20:28 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has joined #code |
| 21:20 | | * McMartin solves #127, reads the real solution |
| 21:20 | < McMartin> | It's got a stateless one, but I cannot comprehend it. |
| 21:20 | < McMartin> | It's a single list comprehension with ten lines of generators and conditions, and uses some syntactic constructs I didn't know Haskell even had. |
| 21:33 | < Rhamphoryncus> | Bah. That's just storing the state on the stack and in temporaries |
| 21:38 | < McMartin> | "stateless" means "directly mutable state". |
| 21:38 | < McMartin> | SSO with reuse of physical memory locations in compiled code still counts as stateless even if it compiles to an identical binary. |
| 21:54 | | crem [moo@Nightstar-8ca3eea7.adsl.mgts.by] has quit [Client closed the connection] |
| 22:00 | < AbuDhabi> | What is the purpose of Reset buttons in online forms? |
| 22:01 | <@TheWatcher> | To reset the contents of the form fields to the value specified in the value attributes when the form was sent |
| 22:02 | <@TheWatcher> | When the fields are empty initially, it's kinda pointless. But sometimes fields come prefilled, and resetting back to those values can be useful sometimes. |
| 22:02 | < AbuDhabi> | Wow. |
| 22:02 | < AbuDhabi> | And here I thought it was just there to annoy users. |
| 22:03 | < AbuDhabi> | Maybe it's its placement. |
| 23:06 | | Thaqui [Thaqui@27B34E.D54D49.F53FA1.6A113C] has quit [Connection closed] |
| --- Log closed Sat Dec 05 00:00:01 2009 |