code logs -> 2008 -> Mon, 18 Aug 2008< code.20080817.log - code.20080819.log >
--- Log opened Mon Aug 18 00:00:20 2008
--- Day changed Mon Aug 18 2008
00:00 You're now known as TheWatcher[T-2]
00:03 You're now known as TheWatcher[zZzZ]
00:07
< Shoukanjuu>
Must...get 13b/specter ikaruga model kit...
00:07
< Shoukanjuu>
I could shell out for two of them, but then we'd still need three more buyers!
00:10
<@McMartin>
What are you producing here?
00:10
< Shoukanjuu>
I'm not actually producing anything of that sort. If you remember, a while ago, I posted a link to a model kit
00:11
< Shoukanjuu>
They need ten orders before they'll produce them
00:11
< Shoukanjuu>
with my order, they need 4 more
00:12
<@McMartin>
Aha.
00:12
< Shoukanjuu>
I could order two, so that when I finish, I have one of each affinity
00:12
< Shoukanjuu>
But still, there's three more people would need to order
00:17 Shoukanjuu [~Shoukanju@Nightstar-18822.dhcp.embarqhsd.net] has quit [Connection reset by peer]
00:21 Shoukanjuu [~Shoukanju@Nightstar-18822.dhcp.embarqhsd.net] has joined #code
00:42 AnnoDomini [AnnoDomini@Nightstar-28950.neoplus.adsl.tpnet.pl] has quit [Quit: Quem quer o garfo?]
02:37 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has quit [Quit: Z?]
02:42 Thaqui [~Thaqui@Nightstar-11452.jetstream.xtra.co.nz] has joined #code
02:42 mode/#code [+o Thaqui] by ChanServ
03:34 Kazriko [~kaz@Nightstar-26352.gdj-co.client.bresnan.net] has quit [Ping Timeout]
03:36 Kazriko [~kaz@Nightstar-26352.gdj-co.client.bresnan.net] has joined #code
03:36 mode/#code [+o Kazriko] by ChanServ
04:52
< Consul>
My closet is its own implementation of bubble sort.
04:52
< Shoukanjuu>
\o/
04:52
< Consul>
The stuff I use most stays to one side.
04:53
< Shoukanjuu>
My closet needs a defrag.
04:53
< Consul>
As the seasons change and the clothes I choose change, it resorts itself.
04:53
<@McMartin>
Is that a Bubble sort or an Exchange Sort?
04:53
<@McMartin>
... actually, I suppose it would actually be an Insertion Sort.
04:53
<@McMartin>
You're taking stuff from the middle and putting it on the end.
04:54
<@McMartin>
(If I read your technique properly)
04:54
< Consul>
Well, I'm taking stuff I want, which could be from anywhere, and after I wear and launder it, it ends up at the end.
04:54
< Consul>
I guess a selection sort, then?
04:55
<@McMartin>
Yeah, it's a little tricky becuase it's a random-access list, really.
04:55
< Shoukanjuu>
I don't like miixmatching stuff. I'll wear either shorts or pants for the entire season.
04:55
< Shoukanjuu>
Usually pants, and usually year round >_>
04:55
<@McMartin>
Selection may be better than Insertion because insertion classically is "take the head and put it in the right place; repeat"
04:56
< Consul>
It was just a dumb off-hand observation. :-)
04:56
< Consul>
Depending on the season, the order of my clothes changes. :-)
04:58
< Shoukanjuu>
...I was just asked what google's main site is.
04:58
<@McMartin>
www.yahoo.com, of course
04:58
< Consul>
Oh, I was going to say askjeeves.com
04:58
<@McMartin>
Aren't they ask.com now?
04:59
< Consul>
Is that one even still around?
04:59
< Consul>
Oh, I suppose they are.
04:59
< Consul>
I used it a few times, and decided I didn't like it.
04:59
< Shoukanjuu>
I told him to google it. I couldn't think of a witty retort for that kind of thing
04:59
<@McMartin>
I knew several people who worked for them before the Crash.
04:59
< Shoukanjuu>
I hope he was kidding
04:59
<@McMartin>
They eventually got married.
04:59
< Consul>
Hehe
05:00
< Consul>
I liked the concept, of asking a question instead of thinking in terms of keywords.
05:00
<@McMartin>
They also had easter eggs as a result of this.
05:01
< Consul>
"Where can I find the best Egg McMuffin porn?"
05:01
< Shoukanjuu>
...........XD
05:01
< Consul>
I swear, if that exists out there somewhere...
05:01
<@McMartin>
It's not there anymore
05:01
< Shoukanjuu>
So I hear Europeans melt at 80?F
05:01
<@McMartin>
But for a time, asking it the average air speed of an unladen swallow demanded a clarification question "African or European?"
05:01
< Consul>
Ha!
05:02
<@McMartin>
Now it treats it as a normal query, though the "narrow your search" questions are at least things like "Are you suggesting that coconuts migrate?"
05:02
< Consul>
I still wonder if the "swim across the Atlantic" direction from Google Maps was an easter egg.
05:02
<@McMartin>
http://www.ask.com/web?q=What+is+the+average+flight+speed+of+an+unladen+swallow% 3F&search=search&qsrc=0&o=0&l=dir
05:03
< Shoukanjuu>
It's like these models fell off the face of the earth!
05:05 * Consul was playing around with a cubic interpolation function in Python.
05:05
< Consul>
My goal being to figure out a way to make an oscillator based around interpolating between a few defined points.
05:05
<@McMartin>
Ooh
05:06
<@McMartin>
If you're doing C++, and you can be open-source, I believe the MikMod implementation also has cubic interpolators
05:06
<@McMartin>
And those are C so bunging them in likely won't be too hard.
05:06
< Consul>
My other goal is to make a memoryless waveshaper using an interpolated table.
05:06
< Shoukanjuu>
I couldn't find any Ikaruga models in stock anywhere.
05:06
< Consul>
MikMod? A Mod player?
05:07
< Shoukanjuu>
So I request ordered it from hobbyfan, and started making an ikaruga in second life.
05:07
< Shoukanjuu>
Using a very nice render I found online.
05:07
<@McMartin>
Consul: Yup.
05:08
<@McMartin>
Notable on the UQM side for tripling your hardware requirements if you used the cubic~
05:08
< Consul>
Well, I have a feeling that I would be better off trying to figure out how to implement Bezier curve oscillators.
05:08
<@McMartin>
Nod.
05:10
< Consul>
I was hoping my idea of using the interpolation formula would be clever.
05:10
< Consul>
But the interp formula doesn't like non-uniform partitions on the X axis.
05:11
< Consul>
I end up having to fill in extra points on the input table, which kinda defeats the purpose of having a few points define a waveform.
05:13
< Consul>
But I got to play with having Python feed data to gnuplot. :-)
05:14
< Consul>
Turns out to be a pretty effective method of prototyping ideas like this.
05:16 Thaqui [~Thaqui@Nightstar-11452.jetstream.xtra.co.nz] has quit [Ping Timeout]
06:11 Thaqui [~Thaqui@Nightstar-17534.jetstream.xtra.co.nz] has joined #code
06:11 mode/#code [+o Thaqui] by ChanServ
08:03 Reiver [~reaverta@Admin.Nightstar.Net] has quit [Ping Timeout]
08:10 Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code
08:10 mode/#code [+o Reiver] by ChanServ
08:14 AnnoDomini [AnnoDomini@Nightstar-28950.neoplus.adsl.tpnet.pl] has joined #Code
08:15 mode/#code [+o AnnoDomini] by ChanServ
08:28 Consul_ [~darren@Nightstar-2021.dsl.sfldmi.ameritech.net] has joined #code
08:28 Consul [~darren@Nightstar-1916.dsl.sfldmi.ameritech.net] has quit [Operation timed out]
09:24 You're now known as TheWatcher
12:04 Thaqui [~Thaqui@Nightstar-17534.jetstream.xtra.co.nz] has left #code [MORE constitution LESS destitution MORE pros...perity.]
12:24 * TheWatcher gets increasingly tempted to just ditch std::list and write his own version
12:30 * TheWatcher EYES the g++-v4 implementation of std::list
12:31
<@TheWatcher>
no, make that is writing his own
12:33 AnnoDomini [AnnoDomini@Nightstar-28950.neoplus.adsl.tpnet.pl] has quit [Ping Timeout]
12:40 AnnoDomini [AnnoDomini@Nightstar-29118.neoplus.adsl.tpnet.pl] has joined #Code
12:40 mode/#code [+o AnnoDomini] by ChanServ
13:18 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has joined #Code
13:18 mode/#code [+o gnolam] by ChanServ
15:55 C_tiger [~c_wyz@Nightstar-6750.hsd1.wa.comcast.net] has quit [Ping Timeout]
16:01 C_tiger [~c_wyz@Nightstar-4230.hsd1.ca.comcast.net] has joined #code
17:09
<@McMartin>
Oh god
17:09
<@McMartin>
never look at C++ headers
17:09
<@McMartin>
Ever
17:09
<@McMartin>
As a C programmer you may recall them being useful for deducing interfaces
17:09
<@McMartin>
EVER SO LOL NO.
17:09
<@McMartin>
Or perhaps ";_; no"
17:10
<@TheWatcher>
Oh, I can read the things... thant's actually rather the problem
17:12
<@McMartin>
Ignoring the fact that it bends over backwards for people who don't want to use new or delete, what exactly is wrong with it?
17:13
<@TheWatcher>
If I purposefully set out to make a linked list class as wasteful as possible in both storage space and operations, I suspect I might be able to top it.
17:14
<@TheWatcher>
It sticks to the spec! Things that should be done in constant time are, things that should be done in linear time are...
17:15
<@TheWatcher>
I can just think of half a dozen ways to make it use less memory, less instructions, and do everything I need.
17:16
<@TheWatcher>
(which, I feely admit, is not the entire spec... but then, I'm not proposing replacing std::list for everyone so I frankly couldn't care about the STL spec)
17:29 Attilla [~The.Attil@92.0.29.ns-20195] has quit [Ping Timeout]
17:31 Attilla [~The.Attil@92.0.29.ns-20195] has joined #code
17:31 mode/#code [+o Attilla] by ChanServ
17:31
<@ToxicFrog>
TheWatcher: so...your complaint is that a specification-matching implementation of a linked list does, in fact, match the spec?
17:36
<@McMartin>
Actually a common complaint regarding C++~
17:38
< Consul_>
Well, I got my interpolated table algorithm working in Python, but it's very hackish and brute-force.
17:38 Consul_ is now known as Consul
17:39
< Consul>
But I suppose that's a decent first step.
17:41
<@McMartin>
First, make it work, then make it right, then make it fast.
17:47
< Consul>
Can I do something like: float* waveshaper = table[500]; -- and set waveshaper to the halfway point of a table so I can refer to it like this: waveshaper[-500] to waveshaper[500]?
17:47
<@ToxicFrog>
float * waveshaper = &table[500]
17:48
< Consul>
Well, float* waveshaper = &table[500]
17:48
< Consul>
Yeah, with ToxicFrog said. :-)
17:48
< Consul>
what
17:48
< Consul>
Argh!
17:48
<@ToxicFrog>
Remember that table[500] is *(table+500)
17:48
<@ToxicFrog>
You can do that, but I'm not sure if the behaviour of negative array indices is well defined
17:49
< Consul>
It's just that transfer function tables in the real world go from a negative to a positive value in range.
17:50
< Consul>
And it would be easier to do that than to set a center offset and add that in every time it's needed, which is what I'm doing in my Python code.
17:52
<@TheWatcher>
TF: no, that the spec doesn't say anything about the implementation /not sucking/
17:52
<@TheWatcher>
¬¬
18:07 AnnoDomini is now known as Steven
18:13 Attilla [~The.Attil@92.0.29.ns-20195] has quit [Ping Timeout]
18:13 Attilla [~The.Attil@92.0.29.ns-20195] has joined #code
18:13 mode/#code [+o Attilla] by ChanServ
18:24
< Consul>
The Interblags seem to think that negative array indices are fine as long as I don't fall outside the array allocation.
18:25
< Consul>
Doing the table lookup that way would remove a number of additions.
18:28
< Consul>
But given that I'm in C++ and I have vectors at my disposal, I might be able to create a different solution.
18:29
< Consul>
On the other hand, this is inner-loop DSP stuff that should be made as fast as possible.
18:30
< Consul>
Cubic interpolation on the inner loop. Welcome to the funhouse!
18:31
<@TheWatcher>
So, how well do you know assembler?~
18:31
< Consul>
I used to program assembler on a 6510. :-)
18:31
< Consul>
Back when there was no single-cycle multiply.
18:32
< Consul>
That was....
18:32
< Consul>
20 years ago.
18:34
< Consul>
Oh my God, I really am 32.
18:35 * Consul goes to dig hide in a hole now.
18:35
< Consul>
err
18:35
< Consul>
Well, you get the idea. :-/
18:49
< Consul>
Well, the interpolator function is purely additions, subtractions, and multiplications. If ever I had an opportunity to try my hand at x86 assembler, now would be a good one.
18:54
< Consul>
That's for a little later though. For now, back to the dreaded preset manager, which has now changed yet again to use SQLite3 as a back end.
18:54
< Consul>
I think this can work well this time, though.
19:09
<@gnolam>
Consul: now you've come of age?
19:09
< Consul>
No, I just realized I'm 32 and have accomplished fuck-all with my life.
19:09 * gnolam was making an obscure music reference.
19:10
< Consul>
Oh.
19:10
< Consul>
Sorry, I missed that one.
19:10
< Consul>
In fact, I'm still missing it... :-/
19:10
<@Steven>
I once programmed a multiplication subroutine for a Motorola 68HC05 processor.
19:11
<@gnolam>
o/~ Welcome to the house of fun
19:11
<@gnolam>
Now I've come of age
19:11
<@gnolam>
Welcome to the house of fun
19:11
<@gnolam>
Welcome to some DSP
19:11
<@gnolam>
Interpolations, yay
19:11
<@gnolam>
Welcome to the house of (fun) o/~
19:11
<@gnolam>
Or something.
19:12
< Consul>
Oh!
19:12
<@gnolam>
Consul: http://www.youtube.com/watch?v=hcE0M5GnhHw
19:12
< Consul>
That is complete and utter Madness, that is.
19:12
< Consul>
:-P
19:25 C_tiger [~c_wyz@Nightstar-4230.hsd1.ca.comcast.net] has quit [Quit: And away she goes!]
19:33 C_tiger [~c_wyz@Nightstar-4230.hsd1.ca.comcast.net] has joined #code
20:00 Shoukanjuu [~Shoukanju@Nightstar-18822.dhcp.embarqhsd.net] has quit [Quit: Shoukanjuu]
20:29
<@ToxicFrog>
McMartin: coda to the VirtualBox experiments: the net claims that win98 runs in protected mode. If so, whatever's causing the performance hit is not real mode emulation.
20:30
<@ToxicFrog>
Possibly it's just that it's spending a lot of time in ring 0, and doing weird things that the VB recompiler can't deal with - in which case the fact that it can't be used with Pacifica means VB doesn't have an efficient way to deal with that.
21:00 Steven is now known as AnnoDomini
22:24 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has quit [Ping Timeout]
22:24 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has joined #Code
22:24 mode/#code [+o gnolam] by ChanServ
23:03 C_tiger [~c_wyz@Nightstar-4230.hsd1.ca.comcast.net] has quit [Ping Timeout]
23:16 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has quit [Quit: Restart.]
23:17 gnolam [lenin@Nightstar-2037.A163.cust.bahnhof.se] has joined #Code
23:18 mode/#code [+o gnolam] by ChanServ
23:25 AnnoDomini [AnnoDomini@Nightstar-29118.neoplus.adsl.tpnet.pl] has quit [Quit: ...and when Mr. Bigglesworth gets upset, people die!]
23:28 You're now known as TheWatcher[T-2]
23:46 You're now known as TheWatcher[zZzZ]
--- Log closed Tue Aug 19 00:00:08 2008
code logs -> 2008 -> Mon, 18 Aug 2008< code.20080817.log - code.20080819.log >