code logs -> 2012 -> Mon, 16 Jan 2012< code.20120115.log - code.20120117.log >
--- Log opened Mon Jan 16 00:00:19 2012
00:00 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code
00:16 You're now known as TheWatcher[T-2]
00:19 You're now known as TheWatcher[zZzZ]
00:29 io|PACKERS is now known as io|BLARGH
00:29 Vash [Vash@Nightstar-8697fea9.wlfrct.sbcglobal.net] has joined #code
00:29 mode/#code [+o Vash] by ChanServ
00:47
< maoranma>
celticminstrel: Mac estas malbone. PC estas pli bone.
02:01 Attilla [Obsolete@Nightstar-019eb622.as43234.net] has quit [[NS] Quit: ]
02:19 io|BLARGH is now known as iospace
02:26 gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Z?]
02:27 Vash [Vash@Nightstar-8697fea9.wlfrct.sbcglobal.net] has quit [[NS] Quit: I <3craft Vorn!]
02:42
< cpux>
It's okay. I'm still trying to learn how to type on an international keyboard. For the sake of communicating in weird inter-languages with colleagues in offices in Europe.
02:44
< maoranma>
Make them all learn Esperanto
03:01 maoranma [maoranma@Nightstar-411b03a4.pools.spcsdns.net] has quit [Connection reset by peer]
03:01
<@Tamber>
So you can /all/ be frustrated to no end!
03:01 maoranma [maoranma@Nightstar-411b03a4.pools.spcsdns.net] has joined #code
03:04
< celticminstrel>
...no, maoranma
03:09
< maoranma>
celticminstrel: Ne? ?u ne?
03:11
< celticminstrel>
No.
03:12
< maoranma>
Sed... Esperanto estas amuzade!
03:21
< cpux>
No it isn't.
03:23
< maoranma>
Pfft, ?ojomortigisto.
03:52 * Kazriko tries to decypher that one.
04:02 Kindamoody[zZz] is now known as Kindamoody
04:11
< maoranma>
Joyslayer
04:11
< maoranma>
Killjoy.
04:24 Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has joined #code
04:32
< Eri|STECSfJLA>
Is there a way to force google to respect case?
04:32 Eri|STECSfJLA is now known as Eri
04:32
< Eri>
Like, I've got some spec. sheets in front of me that refer to charging current as CmA
04:33
< Eri>
Searching for that returns a lot about Certified Management Accountants, but nothing about my usage
04:35
<@himi>
Try googling 'charging current CmA'
04:35
< Eri>
Yeah, I used something like that and got a hit
04:41
< celticminstrel>
Does Google Verbatim respect case?
04:42
< Eri>
That's the special option hidden in a corner of the search options? Let me check
04:42 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Ping timeout: 121 seconds]
04:43
< Eri>
Hmm, I can't even find the option
04:44 * Eri does a search on how to do a search
04:45
< Eri>
Nope. It still doesn't care about case
04:48 eckse_ [eckse@Nightstar-f0ed4a7a.dsl.sentex.ca] has joined #code
04:48 eckse [eckse@Nightstar-435a7b09.dsl.sentex.ca] has quit [Ping timeout: 121 seconds]
04:52 Kindamoody is now known as Kindamoody|out
04:54 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code
04:56
< Eri>
Ha! Shows how well I was reading that document. Turns out I had the answer in front of me the whole time, in the definitions section of this report
05:00 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Ping timeout: 121 seconds]
05:11 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code
05:21 Stalker [Z@Nightstar-3602cf5a.cust.comxnet.dk] has quit [[NS] Quit: Into the hole again, we hurried along our way, into a once-glorious garden now seeped in dark decay.]
05:27 cpux [cpux@Nightstar-c5874a39.dyn.optonline.net] has quit [[NS] Quit: Well, most things get better when I kick them!]
06:26 celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [Client closed the connection]
06:37 eckse_ [eckse@Nightstar-f0ed4a7a.dsl.sentex.ca] has quit [Client closed the connection]
06:46 Eri [Eri@Nightstar-3e5deec3.gv.shawcable.net] has quit [Ping timeout: 121 seconds]
06:49 himi [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds]
06:57 Derakon is now known as Derakon[AFK]
06:59 Eri [Eri@Nightstar-3e5deec3.gv.shawcable.net] has joined #code
09:45 You're now known as TheWatcher
11:56 Attilla [Obsolete@Nightstar-019eb622.as43234.net] has joined #code
13:24 gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code
13:43 Attilla_ [Obsolete@Nightstar-85c2e515.as43234.net] has joined #code
13:44 Attilla [Obsolete@Nightstar-019eb622.as43234.net] has quit [Ping timeout: 121 seconds]
14:01 Attilla_ is now known as Attilla
14:52 celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has joined #code
14:55 * TheWatcher stabs opera while yelling "When I tell you to reload the fucking stylesheet, I mean 'reload the fucking stylesheet', not 'blithely use the one in the sodding cache'"
14:55
< gnolam>
It still does that?
14:57
<@jerith>
Add more meta keys to your "refresh me harder" chord?
14:59
<@TheWatcher>
jerith: I've tried everything up to, and including, all ctrl, shift, and alt keys at once. It still refuses to reload it unless I clear everything
14:59
<@TheWatcher>
It was actually behaving well for a while, there.
15:00
<@jerith>
TheWatcher: Kill it with fire.
15:02
<@TheWatcher>
The vastly irritating part? Shift-reload will force it to send another request for the non-existent favicon.ico file, but it doesn't even send a request for the .css to possibly check whether it might have expired.
15:08
<@jerith>
:-(
15:12<~Vornicus> that seems unawesome.
15:30
<@jerith>
Quite.
15:32
<@TheWatcher>
That's certainly one way to put it
15:48 Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Client exited]
16:36 Vash [Vash@Nightstar-8697fea9.wlfrct.sbcglobal.net] has joined #code
16:36 mode/#code [+o Vash] by ChanServ
16:57 * TheWatcher args, hairpulls at this code
16:57
<@TheWatcher>
how is this being called in a loop, there is no sodding loop that calls these functions?!
17:01 Kindamoody|out is now known as Kindamoody
17:24 * Vornicus pokaes vaguely at code.
17:26
< maoranma>
Uh, is your style sheet cached by the server? I had an issue like that once with some forum software
17:28
< maoranma>
gonna go check my student portal for some python books
17:28
<@TheWatcher>
Nope.
17:28
< maoranma>
if you have any recommendations, that would be awesome
17:29 * Vornicus learned python from the tutorials.
17:30
< maoranma>
http://docs.python.org/tutorial/ <- ?
17:31<~Vornicus> that, yes
17:34
< maoranma>
So many Python books
17:36
<@TheWatcher>
If it helps narrow things down, I always tend to lean towards O'Reilly for such things.
17:39
< maoranma>
Yea, books24x7 has Beginning Python: Using Python 2.6 and 3.1
17:58 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has joined #code
18:01
< gnolam>
Back when I was still suckered into doing web stuff, I held off testing anything in Opera until I'd tested everything else and it was time to upload it to the server.
18:01
< gnolam>
Just because of the overly aggressive caching.
18:02
< gnolam>
Never found a reliable way around it.
18:05 Kindamoody is now known as Kindamoody|cooking
18:05 * Vornicus examines his plan file. has no idea what he's up to in some of these places.
18:05
< gnolam>
... now this is just a wonderfully bizarre solution: http://www.taswegian.com/MOSCOW/rasa.html
18:05
< gnolam>
Want!
18:13
<@TheWatcher>
That is truly odd
18:14
<@TheWatcher>
Ingenious, but odd
18:16 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has quit [Ping timeout: 121 seconds]
18:20 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has joined #code
18:22
< jeroid>
gnolam: Does that predate nixie tubes?
18:26 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has quit [Client closed the connection]
18:26 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has joined #code
18:30 * Vornicus fiddles with euler 60 instead. Apparently doing pregeneration is actually worse than useless.
18:31
< jeroid>
I must do some Euler again.
18:31<~Vornicus> oh, fuckyeah
18:31<~Vornicus> Memory error!
18:32 * jeroid offers Vornicus an Alzheimers.
18:33
< gnolam>
jeroid: nope :)
18:34
< gnolam>
That thing is "ca 1970-1973".
18:34
< gnolam>
Which just adds to the charm.
18:37
< jeroid>
Didn't they already have seven segment display by then?
18:37
< jeroid>
And VFD?
18:39<~Vornicus> gnah, it feels like there should be a better way to do this.
18:39
< jeroid>
There is.
18:39
< jeroid>
But I'm not going to tell you how.
18:39 * Vornicus patpats jeroid
18:46
< gnolam>
Exactly!
18:47
< gnolam>
But someone still went "Let's go with the spinning cylinder. I want to see if it works."
18:51 Kindamoody|cooking is now known as Kindamoody
18:53
< gnolam>
(That's the neat thing about Soviet stuff. With no profit in mind, they could try some really weird things.)
18:55
< gnolam>
https://www.youtube.com/watch?v=qJcOq5sLxPo <- huh. So that's what the guy who made "Ports of Call" is up to these days.
18:58<~Vornicus> awesome.
18:58 Kindamoody is now known as Kindamoody[zZz]
18:59
< jeroid>
Vorn: I have a task for you.
18:59<~Vornicus> ?
18:59
< jeroid>
Find the curvature of a B-spline at a given point.
19:00<~Vornicus> Ah, a good one.
19:01
< jeroid>
And then integrate it over the range affected by a control point.
19:01<~Vornicus> Okay, let's see.
19:01
< maoranma>
You know, I could write something that works like a class in mIRC, but I've literally never done so
19:02
< maoranma>
Well, aliases sort of work like that I guess
19:03 * Vornicus does some hunting, finds a function for curvature of a parametric curve.
19:03
< maoranma>
But there is a way to turn an alias into an identifier
19:04 * Vornicus then hunts up the function for tangential angle. That's important too.
19:04
< maoranma>
Sounds important I tell you whut
19:05
< jeroid>
maoranma: Why would you want to?
19:05
< maoranma>
I wouldn't, just an observation
19:06
< jeroid>
(That's a question, not a dismissal.)
19:06
< maoranma>
Well, I suppose if I needed to I could. I guess most of my projects in mIRC scripting have been too small to deal with it
19:07<~Vornicus> woo, (x'y'' - y'x'') / (x'^2 + y'^2)^(3/2)
19:07
< jeroid>
Vorn: I found one that used first and second derivatives.
19:07
< jeroid>
Yeah, that one.
19:07<~Vornicus> RIght, got that far.
19:08
< jeroid>
I'm not sure I've correctly differentiated the spline.
19:08
< maoranma>
Are we past reticulating them now?
19:08<~Vornicus> Oh yes.
19:09<~Vornicus> Okay, I have to figure out the spline's function. We can split it into piecewise beziers, right, have you succeeded at that?
19:09
< jeroid>
The only method I found skipped some steps I can't work out, and assumes a clamped spline.
19:10
< jeroid>
I've done that, but I'm not sure we want to do the smoothing on the Beziers.
19:12
< jeroid>
I'm also using a shortcut that only works for quadratic splines.
19:12<~Vornicus> What you might end up doing - and this is why this step takes a long time - is dropping to beziers to find the curvature (because finding curvature on beziers is relatively easy) and then hopping back up to the original to do that.
19:13<~Vornicus> er, to adjust points.
19:16
< jeroid>
The shortcut for quadratic splines means that the Bezier control points are at the spline control points.
19:16
< jeroid>
And the Bezier endpoints are at the spline knots.
19:17<~Vornicus> I'm personally still confused at how the spline knots get their locations.
19:18 * Tamber ties knots in Vornicus' splines.
19:18<~Vornicus> Part of this is that I haven't actually done the math and code enough to follow it around.
19:21
< jeroid>
The knots are points in the parametric domain (which is normalised to [0,1]) that divide it into segments which are influenced by different control points.
19:25<~Vornicus> Oh, I see what's going on now.
19:25<~Vornicus> I think.
19:28<~Vornicus> THe knots define points at which different control points come in and out of influence; we use those recursive basis functions to define where each one comes in.
19:29
< jeroid>
You use de Boor's algorithm to get the coordinates for a given input value.
19:29
< jeroid>
Yes.
19:32<~Vornicus> Okay. So, between any two knots I can define a cubic (we're working in cubics, right?).
19:33<~Vornicus> Or rather, a parametric cubic, which is to say that both x(t) and y(t) are cubics.
19:33<~Vornicus> Then /those/ I can do curvature analytically.
19:36 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has quit [Ping timeout: 121 seconds]
19:36<~Vornicus> But I do not think that whatever comes out of there will be integrable directly; that ^3/2 up there doesn't help, let's see what I can do.
19:36
< maoranma>
I wonder if SNES games can display unicode...
19:37
< maoranma>
Might have to modify a font table or something
19:37<~Vornicus> There are some SNES games that have done japanese.
19:37<~Vornicus> You don't get any free fonts on SNES.
19:39
<@iospace>
http://www.regretsy.com/2012/01/16/seven-women-with-stupid-shit-on-their-heads/
19:39
<@iospace>
yes
19:39<~Vornicus> And your RAM and cart size are very small; you get 16 megabytes to play with, combined.
19:39
<@iospace>
that third image is exactly what you think it is
19:40 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has joined #code
19:40 jeroid [jerith@687AAB.5E3E50.F8C097.45AC7B] has quit [Connection closed]
19:40
< maoranma>
I think the pheasant wing one is quite clas*snrk*sy.
19:41 * Vornicus goes back to poking at the spline thing.
19:47
<@ToxicFrog>
maoranma: in practice, SNES games have no concept of "character encoding"; character display is roll-your-own with whatever bitmapped font table the game needs.
19:47
<@ToxicFrog>
Which might be a japanese character set, or a european one, or include weird custom symbols for plot or optimization reasons, or...
19:48
< maoranma>
yea, it's been years since I've done it, so I'm remembering rather slowly
19:48
<@ToxicFrog>
That said, there's nothing (other than memory limitations) stopping you from writing a unicode display engine and character table to run on the SNES.
19:49
< maoranma>
I don't need so much, I'd be adding 6 characters
19:49
< maoranma>
Need to find a working copy of Mario Is Missing
19:53 * Vornicus eyes de boor's algorithm, which is well-explained nowhere.
20:01<~Vornicus> also it would help if I had actual test data!
20:18<~Vornicus> Also it would help if I knew how to animate anything.
20:19
< maoranma>
stick figures?
20:19 * Vornicus patpats maoranma.
20:20
< maoranma>
did I do good?
20:21<~Vornicus> no.
20:21
< maoranma>
aww
20:26 * Vornicus writes .... a class that's a polynomial.
20:26
<@ToxicFrog>
maoranma: what are you trying to do?
20:32 * Vornicus gives it addition, subtraction, multiplication, and (number-only) division tools.
20:43 maoranma [maoranma@Nightstar-411b03a4.pools.spcsdns.net] has quit [Client closed the connection]
20:43 maoranma [maoranma@Nightstar-411b03a4.pools.spcsdns.net] has joined #code
20:44
< maoranma>
wtrf
20:44
< maoranma>
ToxicFrog: I wanted to translate some games into Esperanto
20:50 Vash [Vash@Nightstar-8697fea9.wlfrct.sbcglobal.net] has quit [[NS] Quit: I <3craft Vorn!]
20:52
<@ToxicFrog>
Unicode is way overkill, then
20:52
<@ToxicFrog>
Just replace the character table with the character set you need
20:59
< maoranma>
Yea, like I said, I wouldn't need so much
20:59 * Vornicus kind of wishes for static typing right about now.
21:01
< maoranma>
?? ?? ?? ?? ?? ??
21:01
< maoranma>
That's what I would need to add to do both english and esperanto
21:01
< maoranma>
So I'd have to add twelve characters
21:05<~Vornicus> Man, compared to french, that's a walk in the park.
21:11
<@jerith>
Vornicus: Quadratic, not cubic.
21:12
<@jerith>
(Which is handy, because the TrueType fonts I eventually want to generate are quadratic.)
21:13<~Vornicus> Oh. Well, good, cuz quadratic is actually easier! But I'm still working on the understructure.
21:14<~Vornicus> I can get a ridiculously long way with algebra.
21:14
<@jerith>
Can you differentiate a spline?
21:14<~Vornicus> Yes I Can.
21:15
<@jerith>
I got stuck trying to fiure out how to solve a differential equation on a recursive function space.
21:15<~Vornicus> You build it piecewise, like I'm doing.
21:15
<@jerith>
Assuming I haven't mangled all the terminology.
21:15<~Vornicus> And yeah, you're blowing up the terminology.
21:16<~Vornicus> You're taking the derivative of a recursively-defined function.
21:16<~Vornicus> also I think this is the first time I've actually used __call__
21:16
<@jerith>
N_i,n(u) = f(u)N_(i,n-1)(u) + g(u)N_(i+i,n-1)(u)
21:17
<@jerith>
When I differentiate that, I get N_(i,n-1)(u) and N'_(i,n-1)(u).
21:18
<@jerith>
Which looks like a differential equation.
21:18<~Vornicus> You should look at the "derivative of a B-Spline" page. It's amazing.
21:19
<@jerith>
Except I couldn't get any further without doing it at specific values of n.
21:19
<@jerith>
I looked at that.
21:19
<@jerith>
That assumes a clamped spline, which I don't have.
21:19 Derakon[AFK] is now known as Derakon
21:19
<@jerith>
So I was trying to derive it understand it better.
21:21
<@jerith>
I have that implemented, but the second derivative gives me a B-spline of degree 0, which I have to special-case in my de Boor implementation.
21:22
<@jerith>
I /think/ what I have makes sense, but I didn't get as far as checking the curvature numbers against the output image to make sure.
21:24 himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code
21:24 mode/#code [+o himi] by ChanServ
21:24
<@jerith>
The problem is that I've had to modify it slightly to make it work with non-clamped splines and I don't understand it well enough to be confident that I haven't broken something by doing that. (Or by not doing something else that I don't know about.)
21:28<~Vornicus> I don't see how you might have needed to modify it for non-clamped splines.
21:28<~Vornicus> The general case on the top is right for all b-splines.
21:29 himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds]
21:29
<@jerith>
Vornicus: Which page are you talking about?
21:30
<@jerith>
I worked from http://www.cs.mtu.edu/~shene/COURSES/cs3621/NOTES/spline/B-spline/bspline-derv.h tml
21:30
<@jerith>
And if you want test data, my repo has a little tools to extract some from the unit tests.
21:31<~Vornicus> Right. Equations 4 and 5 are all you need.
21:31<~Vornicus> Or rather 3 and 4.
21:33<~Vornicus> It gets kind of funny on high multiplicity knots though, you'll find that knots of multiplicity p actually eat values so they're not needed any more.
21:33
<@jerith>
Yes, but then you need to drop knots.
21:34
<@jerith>
I've dropped first and last (as they say), but that leaves me with dangling ends.
21:34
<@jerith>
Which I /think/ is okay, but I don't have the math to confirm.
21:44<~Vornicus> I think it's safe, let me do the math.
21:46
<@jerith>
I think it's safe, because it doesn't actually change the domain.
21:48<~Vornicus> Yeah, that... should be safe.
21:51<~Vornicus> YOu end up with one less control point and one smaller degree; this means you have to drop two knots, and you'll find that you're not looking at two of the knots when you do your calculations: 0 and n.
21:51<~Vornicus> er, m.
21:52 * jerith nods.
21:53 * jerith tries a little Portal from laptop-in-bed (which is unlikely to work very well, but it's worth a shot) and then sleep.
21:53<~Vornicus> Yeah no. Mouse is practically required.
21:54
<@jerith>
Vornicus: I'm reasonable with the trackpad on this thing.
21:54
<@jerith>
I've got almost as much accuracy as with a mouse, but rather less speed.
21:55
<@jerith>
And Portal tends not to be action-hero stuff. More puzzley.
21:55
<@jerith>
If you could explain how that spline differentiation works at some point, I'd appreciate it.
22:05 SchoolPhox [8e684915@Nightstar-36f67fd0.mibbit.com] has joined #code
22:07
< SchoolPhox>
So, riddle me this. I'm trying to avoid errors in the equation x/(x+1) - 1, for large values of x. Will multiplying the first term by the denominator's conjugate avoid that?
22:08
< SchoolPhox>
Err, actually, nevermind. I should have actually simplified my terms, to see that it becomes 1/1 - 1
22:09
< SchoolPhox>
Oh, shoot, I thought I was in math
22:09<~Vornicus> Jesus christ, why am I implementing a significant portion of a computer algebra system here.
22:10
< celticminstrel>
I managed to get through Portal with a trackpad, as I recall.
22:10
< celticminstrel>
Though I don't recall very well... it's possible I broke down and plugged a mouse in.
22:12<~Vornicus> Especially when doing the double fling I needed to aim better than the mouse let me.
22:13<~Vornicus> trackpad*
22:56 You're now known as TheWatcher[T-2]
22:58 You're now known as TheWatcher[zZzZ]
23:01 himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code
23:01 mode/#code [+o himi] by ChanServ
23:10 * jerith gave up on the Portalpaults.
23:11
<@jerith>
Not because of the trackpad. I need a better keyboard angle than laptop-in-bed.
23:15 celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.]
23:46
< maoranma>
Yea, I need some type of lapdeskthing for laptoping in bed
23:52 PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds]
--- Log closed Tue Jan 17 00:00:34 2012
code logs -> 2012 -> Mon, 16 Jan 2012< code.20120115.log - code.20120117.log >

[ Latest log file ]