#code.20090104.log

code logs -> 2009 -> Sun, 04 Jan 2009

< code.20090103.log - code.20090105.log >

--- Log opened Sun Jan 04 00:00:23 2009
00:07		Consul [~consul@Nightstar-2535.dsl.sfldmi.ameritech.net] has joined #code
00:07		mode/#code [+o Consul] by ChanServ
00:08	<@Consul>	Well, fsck is happy about both partitions.
00:09	<@Consul>	So, that brings me back to the question of why making a DVD from Youtube downloads (mp4) with Devede is suddenly locking up my computer, when before it worked fine.
00:10	<@ToxicFrog>	You're sure it's the entire system hanging, not just X?
00:10	<@Consul>	Not even the keyboard works.
00:10	<@Consul>	Num or caps lock won't turn on and off.
00:10	<@ToxicFrog>	Aah.
00:10	<@Consul>	ctrl+alt+backspace is no good.
00:10	<@ToxicFrog>	Nor is ctrl-alt-F1? Yeah, not cool.
00:11	<@Consul>	ctrl+alt+del is no good, either.
00:11	<@Consul>	I didn't try the F1 variant, but I have a feeling I can answer it already.
00:12	<@Consul>	It seems like the BBC is burning through a lot of Doctor Who's lately.
00:12	<@ToxicFrog>	What, are we on the 11th Doctor already?
00:12	<@Consul>	Yeah.
00:13	<@Consul>	But before the series was restarted, they'd last a lot longer than this.
00:13	<@Consul>	It's a very different culture at the Beeb, these days.
00:13	<@ToxicFrog>	Some longer than others, but in general, yes.
00:14		Vornicus [~vorn@Admin.Nightstar.Net] has quit [Ping Timeout]
00:16		* ToxicFrog writes some docs for his SDL binding
00:20		Vornicus [~vorn@Admin.Nightstar.Net] has joined #code
00:20		mode/#code [+o Vornicus] by ChanServ
00:21	<@Serah>	Return of the Vorn.
01:20		McM[SanDiego] is now known as McMartin
01:53		Reiver [~reaverta@Admin.Nightstar.Net] has quit [Quit: I ATENT'T DEAD]
02:04		Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code
02:04		mode/#code [+o Reiver] by ChanServ
02:11		AnnoDomini [~farkoff@Nightstar-28974.neoplus.adsl.tpnet.pl] has quit [Quit: Sometimes you burn with the Art, and other times, the Art burns you.]
02:44		Vornicus [~vorn@Admin.Nightstar.Net] has quit [Quit: ]
03:19		* McMartin fiddles with ocamllex
03:24	<@McMartin>	... or, the problem could be in unrelated code managing the command line.
03:26		Serah [~Z@87.72.35.ns-26506] has quit [Connection reset by peer]
03:32		Serah [~Z@87.72.35.ns-26506] has joined #Code
03:32		mode/#code [+o Serah] by ChanServ
03:37		Derakon [~Derakon@Nightstar-4920.hsd1.ca.comcast.net] has joined #code
03:37		mode/#code [+o Derakon] by ChanServ
03:37		* Derakon successfully sets up a bullet flywheel.
03:39	<@Derakon>	Two spirals spewing out one bullet per frame, with bullets moving at 5 pixels per frame, gets me a steady-state at 158 bullets with no noticeable slowdown.
03:40		* Derakon doubles it, gets some slowdown. Drat.
03:41	<@Derakon>	Gonna have to be more efficient somehow...and I'm not even doing collision detection right now.
03:41	<@McMartin>	Heh
03:41	<@McMartin>	Hm. OK, I guess my real problem is that I'm using the wrong version of OCaml.
03:41	<@Derakon>	Collision detection will be a straightforward circle-point algo.
03:45	<@Derakon>	Okay, four emitters, firing bullets in spirals, while they themselves move in circles?
03:45	<@Derakon>	I can't dodge that.
03:56	<@McMartin>	Heh
03:56	<@McMartin>	finally
03:56	<@McMartin>	Though I think this still only works with the latest version of OCaml, but so it goes.
03:57		* McMartin manages to get an autogenerated lexer that properly tracks file/line/column on each token.
05:02	<@McMartin>	ocamlyacc, however, still can't integrate right with this.
05:20		* Derakon ponders the best way to handle BulletML in Python.
05:20	<@Derakon>	I wonder how efficient it would be to generate a set of anonymous functions that describe the behaviors specified in the document.
05:24	<@McMartin>	Why anonymous?
05:24	<@McMartin>	Why not named?
05:25	<@McMartin>	Heck, why not objects, so that they can have state if they need it
05:25	<@McMartin>	Though really, if it's got __call__, you don't care, right~
05:25	<@Derakon>	Heh.
05:25	<@Derakon>	Anyway, the thought was to avoid getting to deeply into a call stack in the interests of efficiency, but that sounds to me like premature optimization.
05:26	<@McMartin>	Lambdas won't cut it for parameterizations.
05:28	<@Derakon>	Hrm...how do I go about viewing the source on that applet...?
05:28	<@McMartin>	Take the jar and run it through a decompiler, most likely.
05:29	<@Derakon>	Oh, wait, there it is.
05:29	<@Derakon>	http://www.asahi-net.or.jp/~cs8k-cyu/bulletml/bulletml0_21.zip
05:32	<@Derakon>	Oh, goody, the source code was procedurally-generated.
05:33	<@McMartin>	Oy.
05:33	<@Derakon>	43k lines of code? No thanks.
05:33	<@McMartin>	Hum.
05:33	<@Derakon>	And the documentation is a bit brief.
05:33		* McMartin wonders if Python has a "procedurally generate a parser from relaxng &c"
05:33	<@Derakon>	I don't think parsing the ML will be especially hard.
05:33	<@Derakon>	Knowing what to do with the parsed objects is the trick.
05:34	<@McMartin>	Yeah
05:34	<@McMartin>	I don't suppose the source was made with RELAXNGCC
05:34	<@Derakon>	It's Java, so I assume gcc is out.
05:36		* Derakon resolves to just take one of the simpler documents (Psyvariar_X-A_boss_opening) and try to make something that looks similar.
05:36	<@McMartin>	RELAXNGCC is in fact Java.
05:37	<@McMartin>	The thing is that it would have the bits he wrote by hand in it
05:37	<@Derakon>	Ah.
05:37		* McMartin pokes at it himself
05:39	<@McMartin>	I suspect that stuff like BulletImpl.java that actually talks to the applet are the parts you'd be able to loot.
05:39	<@Derakon>	Well, I'm gonna fool about with getting ElementTree working for the nonce.
05:39	<@McMartin>	Let me know how it goes; I'd like to hear about its utility.
05:40	<@McMartin>	Aha.
05:40	<@Derakon>	Will do.
05:40	<@McMartin>	He used "Relaxer" for the procedurally generated stuff
05:40	<@McMartin>	In short, what you'd be using Elementtree for.
05:40	<@Derakon>	Fun times.
05:40	<@McMartin>	That's the XML parsing because Java's XML support is the reference standard.
05:40	<@McMartin>	Read: It sucks.
05:40		* Derakon snickers.
05:44		* McMartin rereads net.sf.blorple.ifiction.Parser's source code, swiftly re-convinces himself of this.
05:45	<@McMartin>	631 lines of actual XML-touching code, with some fairly sophisticated abstractions, and several thousand lines of data representation bullshit
05:45	<@McMartin>	Just to get some decent structures.
05:46	<@McMartin>	Yeah, that would have been much easier with something like Relaxer.
05:49	<@Derakon>	Blorple did its own XML management?
05:49	<@Derakon>	Or it used the Java standard implementation?
05:50	<@McMartin>	It used the Java standard implementation and then reified it as a set of basic structure objects.
05:50	<@McMartin>	So that Blorple proper doesn't have to give a shit about XML, just the Story class and its many-nested components.
05:53	<@McMartin>	Oh, man
05:54		* McMartin summons Steele, Gosling, and Stroustrup to beat the writer of BulletML for his sins against OO.
05:54	<@Derakon>	BulletML the markup, or the implementation?
05:55	<@McMartin>	ActionImpl.java is the one that has what you want
05:55	<@McMartin>	It is done with a chain of "if (x instanceof blah) { ... } else if (x instanceof blah) { ... } ..."
05:55	<@McMartin>	VIRTUAL FUNCTIONS: NOT JUST FOR BREAKFAST ANYMORE
05:56	<@Derakon>	Ew.
05:56	<@McMartin>	(To be fair, since those blah functions may have been autogenerated, he might not have been able to add stuff like that.)
05:57	<@Derakon>	This is 322 lines of basically uncommented Java with badly-chosen variable names.
05:58	<@McMartin>	Still probably easier to read than the raw bytecode~
05:58		* Derakon facepalms.
05:58	<@McMartin>	That's the closest thing to a spec I can see, though.
05:59	<@Derakon>	Yeah, thanks for tracking it down.
06:00	<@McMartin>	I'd go by the minimal reference first, and only go to the source when an action is wildly underspecified instead of merely underspecified.
06:00		* Derakon nods.
06:00	<@McMartin>	(See also: me and iFiction.)
06:01	<@McMartin>	(In which I used the babel tools for the fallback.)
06:01	<@Derakon>	Okay, I have ElementTree parsing the document and getting the right element tags out, at any rate. Now to try accessing labels/contents.
06:02	<@McMartin>	Already Better Than MiniDOM
06:03	<@Derakon>	Here's the tree-print code I threw together: http://paste.ubuntu.com/99428/
06:04	<@Derakon>	Not very interesting, but I figured you might be curious.
06:06	<@jerith>	Derakon: ElementTree's in the stdlib from 2.5 (as etree, iirc).
06:06	<@Derakon>	Ahh?
06:06	<@McMartin>	Well, that just means he just has to change the import statement.
06:07	<@jerith>	Yeah.
06:07	<@Derakon>	"import etree" doesn't work, anyway.
06:07	<@Derakon>	And I'm using 2.5.
06:07	<@McMartin>	It's deeper than that...
06:07	<@jerith>	What I do is try import the stdlib version and fall back to the external library if it's not there..
06:07	<@jerith>	-.
06:08	<@McMartin>	import xml.etree.ElementTree as ET
06:08	<@Derakon>	That works.
06:08	<@Derakon>	How did you figure that out?
06:08	<@McMartin>	I looked at http://docs.python.org/library/
06:08	<@Derakon>	Ah hah.
06:08	<@McMartin>	Alternate answer:
06:08	<@McMartin>	AVATAR
06:09	<@Derakon>	Dangit, Steve!
06:09		* McMartin checks his shirt
06:09	<@McMartin>	I'm definitely not Steve.
06:09	<@Derakon>	I'm safe from cannons, then?
06:09	<@jerith>	Then catch ImportError (or whatever) and look for elementtree.
06:09	<@McMartin>	Can anyone ever truly be safe from cannons?
06:09	<@McMartin>	(import cannon)
06:09		* Derakon facepalms.
06:10	<@jerith>	FLYING ARTILLERY! (import cannon; import antigravity)
06:11	<@McMartin>	"When will your reign of terror end, Death Cannon?"
06:11	<@McMartin>	"If anyone asks, he fell down some stairs. ...into a cannon."
06:12	<@Derakon>	Okay, lessee...I have a <bullet> tag which describes starting properties of a bullet (notably not including position...), an <action> tag that lets me change the properties of bullets, a <fire> tag to instantiate bullets, and a bunch of other things that are contained within those tags.
06:12	<@Derakon>	Actions can also exist independent of bullets, e.g. for a main loop to create initial bullets.
06:13	<@Derakon>	And everything can be named and referred to by name elsewhere in the document.
06:13	<@jerith>	You're firing bullets in XML?
06:13	<@Derakon>	BulletML.
06:14	<@Derakon>	XML for danmaku ("bullet hell" / "manic shooter") games.
06:14	<@jerith>	Hrm.
06:14	<@Derakon>	Basically a markup language for specifying the behaviors of particles.
06:14	<@jerith>	Can we fire bullets into XML instead?
06:14	<@Derakon>	I'm sure it's been done many times.
06:15	<@jerith>	Not by me.
06:15	<@Derakon>	Don't let me stop you, then.
06:15		* jerith resolves to take a printout of the SOAP spec next time he goes to the range.
06:15	<@Derakon>	Anyway, from what I can tell, a useful BulletML document must have either an action tag or a fire taga at the root level
06:16	<@Derakon>	Since otherwise nothing would happen.
06:16	<@jerith>	Shouldn't you need something else to do the firing?
06:16	<@Derakon>	Also, the position of the bullet emitters is not referred to anywhere in the document; that's assumed to be handled by the application.
06:17	<@Derakon>	Think of this as the code that gun turrets execute, Jerith.
06:18	<@Derakon>	E.g. "repeat 100 times: fire a bullet(angle 45 +- rand(5) degrees, speed .5 +- rand(1)); sleep(2)"
06:18	<@jerith>	Ah, right.
06:19	<@jerith>	So this is more ShootML than BulletML.
06:20	<@Derakon>	Um, sure?
06:21		* jerith was thinking of it a a description of a bullet rather than the description of the bullet and the fire control for said bullet.
06:21	<@jerith>	*as
06:21		* Derakon nods.
06:21		* Derakon tries to figure out the right way to handle this whole thing.
06:22	<@jerith>	On first thought, I'd split it into bullets and guns.
06:23	<@Derakon>	Python supports currying, right?
06:23	<@jerith>	Bullets describe things like damage, explosions, proximity fuses, etc.
06:23	<@jerith>	Derakon: Sort of. You have to return a closure from a function, it doesn't happen automagically.
06:24	<@jerith>	Guns describe things like magazine capacity, rate of fire, etc.
06:24	<@Derakon>	What I'm thinking is that I'd have e.g. an action() function which accepts as inputs all the valid contents of an <action> tag, as well as an object to operate on (a bullet, or null in which case it'd return new bullets or sommat).
06:24	<@McMartin>	jerith: BulletML is specifically designed to be a spec language for agents in Bullet Hell games.
06:24	<@Derakon>	So when I parse an <action> tag I invoke action() will all but one of the arguments, and take the resulting function and stick it into a list of functions that I'll need.
06:24	<@jerith>	McMartin: Ah, I lack context then.
06:24	<@McMartin>	Derakon: Python is Weird about closures.
06:25	<@McMartin>	You will want to return objects that do the python equivalent of overriding operator().
06:25	<@Derakon>	So e.g. a simple <bullet> tag might just provide the angle and speed of the bullet; I call bullet(angle, speed), get a function back, and use that function when updating generated bullets.
06:25	<@McMartin>	Because "closures" are not permitted to have writable state.
06:25	<@Derakon>	I don't think I follow.
06:25	<@McMartin>	So, if this were Scheme
06:26	<@McMartin>	You'd return a lambda expression that referred to some values that were passed to the lambda's generator as parameters.
06:26	<@McMartin>	This sort of works in Python, if and only if those things are read-only
06:26	<@McMartin>	Because otherwise there's a lack of clarity about whether you were writing to something in a higher scope or creating a new variable that shadows it, so Python just bans it outright.
06:27	<@McMartin>	The preferred solution is to return an object with fields, thus clarifying what gets written where.
06:27	<@Derakon>	Couldn't just enforce pass-by-value or something, huh?
06:27		* jerith packs up to go to the office for the day.
06:27	<@McMartin>	Derakon: This has nothing to do with passing, and more to do with "what does the name 'x' mean"
06:27	<@Derakon>	Okay, so when I parse a <bullet> tag, I'd create a new BulletFunc object, set its fields (e.g. angle, speed), and Bullet instances would contain BulletFuncs, which are called during Bullet.update?
06:28	<@McMartin>	Ideally, during Bullet.__call__, I'd think
06:28	<@McMartin>	You want these to act like functions, right?
06:28	<@Derakon>	That is, Bullet.update would call BulletFunc.updateBullet or something.
06:28	<@Derakon>	That would be ideal.
06:28	<@Derakon>	I'm not familiar with __call__.
06:28	<@McMartin>	Let me test this in a REPL to make sure I get it right
06:28	<@Derakon>	Ahh. __call__ overloads operator().
06:29	<@McMartin>	Yeah.
06:29	<@Derakon>	Handy, tha t.
06:29	<@McMartin>	But it may act oddly with __init__.
06:29	<@Derakon>	s/tha t/that/
06:29	<@Derakon>	http://www.daniweb.com/forums/thread39004.html
06:29	<@Derakon>	That implies that Python is smart about the context of the call.
06:29	<@McMartin>	It is on reads.
06:29	<@Derakon>	I.e. if you're using the class name, it calls __init__; if you're using an instance, it calls __call__.
06:29	<@McMartin>	Oh
06:29	<@McMartin>	Yeah
06:30	<@McMartin>	>>> class A:
06:30	<@McMartin>	... def __init__(self, x): self.x = x
06:30	<@McMartin>	... def __call__(self, y): return self.x + y
06:30	<@McMartin>	...
06:30	<@McMartin>	>>> a = A(2)
06:30	<@McMartin>	>>> a(3)
06:30	<@McMartin>	5
06:30	<@McMartin>	(Too short to pastebin)
06:30		* Derakon nods.
06:30	<@Derakon>	Okay, that seems workable.
06:30	<@McMartin>	If something isn't paramaterized, like, say, vanish tags, then those probably can be simple functions.
06:30	<@jerith>	Back in a few minutes.
06:30	<@Derakon>	Right.
06:30	<@McMartin>	They can't be lambdas, I don't think, because lambdas are restricted to a single expression, and it can't be a statement.
06:31	<@McMartin>	And besides, you have a name for it: "vanish".
06:32	<@Derakon>	Hm. I'll need an internal field on Actions to indicate that they're "done" so I know to move to the next thing or to force-repeat them.
06:32	<@Derakon>	...agh, this does not interact well with an external main loop.
06:33		Vornicus [~vorn@Admin.Nightstar.Net] has joined #code
06:33		mode/#code [+o Vornicus] by ChanServ
06:33	<@Derakon>	Every time there's a <wait> tag I'll need to basically set an interrupt in the middle of the function.
06:33	<@McMartin>	Have agents have a countdown of "frames to next action"?
06:33	<@McMartin>	With the wait action tracking to "set this value to something that isn't 1"?
06:33	<@Derakon>	An action could be <wait>2</wait><fire>foo</fire><wait>2</wait><fire>bar</fire>
06:34	<@McMartin>	"action" is really a program, it sounds like.
06:34	<@Derakon>	Pretty much, yeah.
06:34	<@McMartin>	What we've been talking about is not for actions, then.
06:34	<@McMartin>	it's for their components: wait, fire, accel, vanish, &c.
06:35	<@Derakon>	I guess <wait> actually means "invoke update/draw cycle once"?
06:35	<@McMartin>	That wasn't how I was imagining the architecture, but you're the architect~
06:35	<@McMartin>	(I sort of flipped the priority there)
06:35		Vornotron [~vorn@Admin.Nightstar.Net] has joined #code
06:35		* Derakon tries removing the <wait> tag from one of the files in the sample implementation, gets code that does actually still draw things.
06:36	<@McMartin>	You can't let individual agents dictate global actions like draw!
06:36	<@Derakon>	I'm trying to figure out how to reconcile this behavior with a simple main loop...I'm not certain it works.
06:36	<@Derakon>	Yeah.
06:36		Vornicus [~vorn@Admin.Nightstar.Net] has quit [Ping Timeout]
06:36	<@Derakon>	This feels like a situation that calls for threads.
06:37	<@McMartin>	The simple main loop tells each agent "It's been a tick. Do your thing." Agents do action processing, probably by waiting for their trigger and if it hits zero, invoking their next sub-action.
06:37	< Vornotron>	why are we using XML for a tick pattern.
06:37	<@McMartin>	If that action was "wait", their trigger is delayed.
06:37	<@McMartin>	Vornotron: We are not. We're using it because it already exists and people use it.
06:37	<@McMartin>	It's for specifying scripted firing pattern AIs for bullet hells.
06:37	<@Derakon>	I'm basically trying to make a Python implementation.
06:37	< Vornotron>	Oh.
06:38	<@McMartin>	Anyway, this is synchronized enough that I don't think threads are strictly necessary.
06:38	<@Derakon>	McM: my issue is how to track where in the action processing an agent is.
06:38	<@McMartin>	Give agents state.
06:38	<@Derakon>	So say an action's code says "fire a bullet, wait a bit, fire a different bullet, wait a bit."
06:38	<@Derakon>	I call the action, it fires the first bullet, hits the first wait statement, returns.
06:38		Vornotron [~vorn@Admin.Nightstar.Net] has quit [Connection reset by peer]
06:39	<@Derakon>	I call it again, it needs to say "Okay, I'm partway into the first wait; I'm not done yet; return."
06:39	<@Derakon>	I call it again, it needs to say "Okay, I'm done with the first wait, now where was I? Okay, fire second bullet, time to wait, so return."
06:39	<@McMartin>	... what about this do you think is tricky?
06:39		Vornicus [~vorn@Admin.Nightstar.Net] has joined #code
06:39		mode/#code [+o Vornicus] by ChanServ
06:39	<@McMartin>	I mean, you've essentally specified an algorithm right there.
06:40	<@McMartin>	The only thing missing is the variable names.
06:40	<@Derakon>	Tracking where in the algo I am.
06:40	<@McMartin>	Right.
06:40	<@McMartin>	So
06:40	<@Derakon>	Do I split out everything between wait statements into separate functions or something?
06:40	<@McMartin>	Our agent has some list: [fire(foo), wait(2), fire(bar), wait(2)]
06:40	<@McMartin>	We have an integer that says where in that list we are.
06:40	<@McMartin>	We have some other integer that says how many times we have to be woken up before we resume processing.
06:40	<@Derakon>	And a dict or something that contains state for each position in the list.
06:41	<@McMartin>	No, because these are callables, right?
06:41	<@Derakon>	Okay, that should work.
06:41	<@McMartin>	The state is living there in "fire(foo)"
06:41	<@Derakon>	Ahh, right.
06:41	<@Derakon>	It's the "break the action code into segments" bit that I was missing, really.
06:41	<@McMartin>	Aha
06:42	<@McMartin>	So yeah, main loop then becomes "wake up each agent once; draw."
06:42	<@Derakon>	Right.
06:42	<@McMartin>	For added fun, the player can be an agent too, but his update involves checking the input state.
06:42	<@McMartin>	(This is effectively what Sable does)
06:42	<@Derakon>	Okay, I need to write this all down since it's getting a bit late.
06:43	<@Derakon>	I've already basically written the player code, so I don't feel the need to muck with it ATM.
07:06		* jerith is back.
07:08		* Derakon sings Vim's praises for search&replace.
07:08	<@Derakon>	Makes changing < to < and > to > much easier.
07:08	<@Derakon>	Anyway, made a big LJ post to my gamedev filter.
07:20	<@Vornicus>	...okay, bulletml?
07:21	<@Vornicus>	/fucking awesome/
07:21	<@Derakon>	Heh.
07:21		* Vornicus is watching the demos, and they are incredible.
07:30	<@Derakon>	To be fair, those are just implementations in BulletML of patterns taken from various games.
07:30	<@Vornicus>	yes.
07:30	<@Vornicus>	But some of these I did not expect you to be able to do.
07:31	<@Vornicus>	Now that I haveseen that they are possible, I agree that bulletml is the way to go.
07:31	<@Vornicus>	(The backburst one is my favorite)
07:34	<@jerith>	Der: Where do I find this post?
07:36	<@Derakon>	http://derakon.livejournal.com/306123.html?#cutid1
07:36	<@Derakon>	I don't think I have you friended, though.
07:36	<@Vornicus>	jerith!
07:37	<@Derakon>	Okay, you should be good to go.
07:39	<@jerith>	Cool, thanks.
07:40		* Derakon determines that in the example implementation, the speed and direction on the <bullet> tag are overridden by the speed and direction in the <fire> tag.
07:40	<@Derakon>	Okay, bedtime. T-2, all.
07:43		Derakon is now known as Derakon[AFK]
08:40		KBot [AnnoDomini@Nightstar-84.neoplus.adsl.tpnet.pl] has joined #Code
08:41		KarmaBot [AnnoDomini@Nightstar-28974.neoplus.adsl.tpnet.pl] has quit [Ping Timeout]
08:43		KBot is now known as KarmaBot
10:37		AnnoDomini [~farkoff@Nightstar-84.neoplus.adsl.tpnet.pl] has joined #Code
10:37		mode/#code [+o AnnoDomini] by ChanServ
10:47		Vornicus [~vorn@Admin.Nightstar.Net] has quit [Quit: ]
14:32		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has joined #Code
15:01		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has quit [Ping Timeout]
15:02		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has joined #Code
15:55		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has quit [Ping Timeout]
15:56		C_tiger_remote is now known as C_tiger
15:57		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has joined #Code
16:00		gnolaptop [~lenin@Nightstar-1449.dialup.nmt.net] has quit [Ping Timeout]
16:03		gnolaptop [~lenin@Nightstar-1583.dialup.nmt.net] has joined #Code
16:08		gnolaptop [~lenin@Nightstar-1583.dialup.nmt.net] has quit [Ping Timeout]
16:14		gnolaptop [~lenin@Nightstar-1844.dialup.nmt.net] has joined #Code
16:18		Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code
16:18		mode/#code [+o Attilla] by ChanServ
16:28		Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Ping Timeout]
16:38		Derakon[AFK] is now known as Derakon
16:58		gnolaptop [~lenin@Nightstar-1844.dialup.nmt.net] has quit [Ping Timeout]
17:31		gnolaptop [~lenin@91.149.34.ns-11340] has joined #Code
17:39		gnolaptop [~lenin@91.149.34.ns-11340] has quit [Ping Timeout]
17:47		gnolaptop [~lenin@91.149.34.ns-11340] has joined #Code
18:00		gnolaptop [~lenin@91.149.34.ns-11340] has quit [Ping Timeout]
18:02		gnolaptop [~lenin@91.149.34.ns-11340] has joined #Code
18:03		gnolaptop [~lenin@91.149.34.ns-11340] has quit [Quit: So far, I'm not liking 2009 one bit.]
19:09		gnolam [lenin@Nightstar-1382.A163.priv.bahnhof.se] has joined #Code
19:09		mode/#code [+o gnolam] by ChanServ
20:57		Vornicus [~vorn@Admin.Nightstar.Net] has joined #code
20:57		mode/#code [+o Vornicus] by ChanServ
21:06		Vornicus is now known as Finerty
21:13		Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code
21:13		mode/#code [+o Attilla] by ChanServ
21:14	<@Consul>	Should I be ashamed to admit I'm enjoying these MIT lectures on differential equations on YouTube?
21:14	<@Finerty>	where?
21:14	<@Consul>	Youtube
21:14	<@Consul>	User MIT
21:14	<@Finerty>	Sweet.
21:14	<@Consul>	They have a playlist for all 32 lectures, ~50 mins each.
21:15	<@Consul>	Creative Commons license, so they can probably be downloaded as a torrent somewhere.
21:16	<@Consul>	http://ocw.mit.edu/OcwWeb/Mathematics/18-03Spring-2006/CourseHome/index.htm
21:16		Finerty is now known as Vornicus
21:16		Vornicus is now known as Finerty
21:16	< simontwo>	Consul, enjoying? yes!
21:17	<@Consul>	They did two things right. 1) The professor is actually a good speaker who can keep you engaged. 2) The videography and sound are well-done and interesting.
21:17	< simontwo>	Consul, I'm currently viewing the ones in Linear Algebra because my own book suddenly got a lot harder when skipping two weeks of lectures.
21:17	<@Consul>	Oh, and 3) they didn't cut out any footage when something quirky happens.
21:18	< simontwo>	Consul, okay. the one on Linear Algebra is constantly asking the audience rhetorical questions (i.e. I don't think he cares what they answer, but still he waits for them to say something before continuing).
21:18	<@Consul>	http://ocw.mit.edu/OcwWeb/Mathematics/18-03Spring-2006/VideoLectures/index.htm -- Video downloads, though no torrent file I can see.
21:18	<@Consul>	simontwo: Oh, well, I've only been watching the diff-eq ones. I finished linear algebra, with an A I might add. :-)
21:18	<@Consul>	I'm starting diff-eq in a week, and I thought I'd get a head-start.
21:19	< simontwo>	Consul, sounds sensible.
21:20	<@Consul>	Eh, guess I'll just keep watching them on Youtube. I was hoping for a torrent.
21:20	< simontwo>	I have a mandatory course after this one called "web programming".
21:24	<@Derakon>	Hmm...all these objects are starting to get a bit hairy.
21:37	<@Derakon>	ElementTree's documentation is a bit lacking in documentation telling me how to get the type of a tag out.
21:38	<@Derakon>	E.g. I want to go from <action label="foo"> to 'action'.
21:38	<@Derakon>	The docs appear to be geared towards writing XML documents, not reading them.
21:46		Attilla_ [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has joined #code
21:47		Attilla [~The.Attil@Nightstar-9469.cdif.cable.ntl.com] has quit [Ping Timeout]
21:48		Attilla_ is now known as Attilla
22:02	<@Derakon>	Ha HA!
22:03	<@Derakon>	Simple BulletML file: PARSED.
22:03	<@Consul>	Derakon: Sweet!
22:03	<@Consul>	I have no idea what that is, but it deserves some congratulations, I'm sure. ;-)
22:04	<@Derakon>	Heh. Check out http://www.asahi-net.or.jp/~cs8k-cyu/bulletml/bulletml_applet_e.html
22:04	<@Derakon>	In short, I'm now handling the Psyvariar_X-A "boss opening" file on that page.
22:04	<@Consul>	Oh, cool! The OpenJDK plugin is working!
22:05	<@Derakon>	...well, okay, the <wait> tag isn't being obeyed. I should fix that.
22:05	<@Consul>	Hey, that's pretty neat...
22:06	<@Finerty>	BulletML: a markup language for describing firing sequences for entities in shooters.
22:06	<@Consul>	Finerty: Thanks...
22:07	<@Derakon>	Okay, I have a bunch of different classes, I want to give each instance of a class a unique ID (regardless of the class type).
22:08	<@Derakon>	I created a variable globalId at the top of the script, so I could set self.id in every init function.
22:08	<@Derakon>	But it won't let me refer to globalId. Says "UnboundLocalError: local variable 'globalId' referenced before assignment"
22:09	<@Derakon>	Ahh. 'global' keyword is my friend.
22:10	<@Finerty>	Yes indeed.
22:10	<@Finerty>	Though I was under the impression that objects got ids from the vm.
22:11	<@Derakon>	Ahh?
22:11	<@Derakon>	Well, it's probably some hexadecimal crud. I know what Action #0 is on each invocation.
22:11	<@Finerty>	Yeah, let me see if I can find that information - object i is the power of is.
22:11	<@Finerty>	object id, rather
22:11	<@Derakon>	That sentence still makes no sense.
22:12	<@Derakon>	"object id is the power of is"?
22:12	<@Finerty>	"is" uses object identity
22:12	<@Finerty>	(the is operator)
22:12	<@Derakon>	Ahh.
22:12		MyCatVerbs [~mycatverb@Nightstar-13709.lurkingfox.co.uk] has quit [Ping Timeout]
22:23	<@Derakon>	There. Wait and repeat now function properly.
22:44		* Derakon creates submunitions.
22:46	<@Derakon>	(Thereby verifying that bullets do run actions contained within them)
23:02		AnnoDomini [~farkoff@Nightstar-84.neoplus.adsl.tpnet.pl] has quit [Quit: There is strength not only in knowing the self, but knowing how to bring it forth in others.]
23:55	<@McMartin>	w00t
--- Log closed Mon Jan 05 00:00:35 2009

code logs -> 2009 -> Sun, 04 Jan 2009

< code.20090103.log - code.20090105.log >