code logs -> 2013 -> Fri, 30 Aug 2013< code.20130829.log - code.20130831.log >
--- Log opened Fri Aug 30 00:00:54 2013
00:29 You're now known as TheWatcher[T-2]
00:31 * McMartin eyes his co-worker
00:31
<&McMartin>
He has written a boot sector that dumps itself to the screen in hex along with the company slogan in ASCII
00:31
<&McMartin>
In 16-bit assembler, because boot sector
00:31
<&McMartin>
That's a fine feat of lightsaber chair fighting.
00:31
<&McMartin>
Maybe we can have this be our T-Shirt~
00:32
<@TheWatcher[T-2]>
...
00:33
<@TheWatcher[T-2]>
Next up: bouncing scrolling text and chip tune music
00:33
<&McMartin>
I would wear the shit out of a T-shirt that did that~
00:33 You're now known as TheWatcher[zZzZ]
00:36
<&McMartin>
Sadly, I actually still remember the segment that holds screen character data~
00:36
<&McMartin>
[b800:0000], in two-byte words, high byte is color/background nybbles (woo 16 colors), low byte is the ASCII code
00:38
<@Reiv>
nibbles?
00:39
<&McMartin>
4 bits.
00:39
<&McMartin>
Half a byte, you see~
00:40
<@Reiv>
snrk
00:40
<@Reiv>
I should have guessed~
00:41
< AnnoDomini>
Now you've done and reminded me of my computer architecture class.
00:43
< AnnoDomini>
Damn it. One of my partitions on this laptop is FREEDOS. And has the assignments for that.
00:44
<&McMartin>
Ha ha, VP of product's reaction is "we need a T-Shirt of this", so maybe we'll get them after all~
00:46
<&Derakon>
If my lab had a T-shirt, it would probably have white noise on the front, and an image of a cell on the back.
00:46
<&Derakon>
(Since we specialize in very-low-light microscopy and post-processing to pull signal out of noise)
00:51
<~Vornicus>
No, it's white noise plus blacklight or glow-in-the-dark cell
00:53
< [R]>
http://john.kaniarz.com/2012/09/designing-attack-damage-formulas.html <-- This guy isn't being very clear here. Why is he dividing health by the damageMultiplier? What's `x`? What's `c`? If `c` is the cap, why is it not constant?
00:54
< [R]>
Also I suspect there's an error: unless `c` is 100, the percentage formula doesn't give a percentage.
00:57 Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving]
01:01 Karono [Karono@9C034E.4BE65E.E00AF8.FDA077] has joined #code
01:37
<&Derakon>
Vorn: well, the point is that images that look to us like they're just noise can have surprising amounts of signal hidden in them if you use the right algorithms.
01:38
<&Derakon>
I've seen useful data reconstructed out of what I would have sworn was pure snow.
01:50 RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has quit [[NS] Quit: Gone.]
01:53 RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has joined #code
01:57 Karono [Karono@9C034E.4BE65E.E00AF8.FDA077] has quit [Ping timeout: 121 seconds]
02:20 Karono [Karono@9C034E.4BE65E.E00AF8.FDA077] has joined #code
02:35 gnolam_ [lenin@Nightstar-f7705974.cust.bredbandsbolaget.se] has joined #code
02:35 gnolam is now known as NSGuest24088
02:35 gnolam_ is now known as gnolam
02:35 mode/#code [+o gnolam] by ChanServ
02:37 NSGuest24088 [lenin@Nightstar-f7705974.cust.bredbandsbolaget.se] has quit [Ping timeout: 121 seconds]
02:44 Vorntastic [Vorn@Nightstar-b8287ec3.sub-70-211-15.myvzw.com] has joined #code
02:44 Vorntastic [Vorn@Nightstar-b8287ec3.sub-70-211-15.myvzw.com] has quit [Connection reset by peer]
02:45 Vorntastic [Vorn@Nightstar-b8287ec3.sub-70-211-15.myvzw.com] has joined #code
02:50
<&McMartin>
http://hylang.org
02:50
<&McMartin>
Relevant to seveal folks in the channel, I imagine
02:55
< [R]>
http://shotgunshuffle.com/wp-content/uploads/2012/10/WTF-Toys-R-Us-4.jpg
03:04
<@Reiv>
... wait, you can do lisp in python?
03:05
<@Reiv>
I thought there were Certain Features a language needed to be able to do the proper tail-end recursion functional programming tomfoolery to begin with
03:11
<@Alek>
you can lisp in any language.
03:16
<&Derakon>
Like Reiv said, languages don't always handle tail-recursion properly.
03:16
<&ToxicFrog>
Reiv: there are various mechanical translations you can do to get TCO semantics even for target architectures (like the JVM) that don't support it.
03:18
<&McMartin>
Clojure does not do them
03:19
<&McMartin>
But they can in fact be done.
03:19
< Vorntastic>
Tail call recursion is nigh-trivially translateable into iteration, which is why the optimization can happen.
03:19
<&McMartin>
Full tail call elimination, however, is only trivial if your compiler uses continuation-passing style as its intermediate representation.
03:19
<&McMartin>
Clojure makes you kind of do this by hand with the trampoline.
03:20
< Vorntastic>
...trampoline?
03:21
<&McMartin>
(In practice, tail recursion is kind of inconvenient and namespace polluting. If your language has it (read: is Scheme specifically), then the named let is much nicer. Clojure's loop-recur is a subset of that.)
03:21
<&McMartin>
http://clojure.github.io/clojure/clojure.core-api.html#clojure.core/trampoline
03:22
<&McMartin>
"algorithms requiring mutual
03:22
<&McMartin>
recursion without stack consumption"
03:22
<&McMartin>
being the use case for fully general TCO.
03:27
<&ToxicFrog>
For non-mutual recursion it has loop/recur.
03:27
<&ToxicFrog>
Scala uses the @tailrec annotation.
03:35 Karono_ [Karono@9C034E.4BE65E.E00AF8.FDA077] has joined #code
03:35 Karono [Karono@9C034E.4BE65E.E00AF8.FDA077] has quit [Ping timeout: 121 seconds]
03:35 Karono_ is now known as Karono
04:08 Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code
04:08 mode/#code [+qo Vornicus Vornicus] by ChanServ
04:09 Vorntastic [Vorn@Nightstar-b8287ec3.sub-70-211-15.myvzw.com] has quit [[NS] Quit: Bye]
04:12 celticminstrel [celticminst@Nightstar-ae361035.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
04:14
<@Reiv>
... so does Haskell do it the messy way, then?
04:15
<&McMartin>
Haskell it Just Works
04:18 Karono [Karono@9C034E.4BE65E.E00AF8.FDA077] has quit [Connection reset by peer]
04:18
<@Reiv>
In ways that the others don't?
04:18
<&ToxicFrog>
It also Just Works in Lua and in most Lisps
04:19
<&ToxicFrog>
Note that all of these have their own compiler and/or VM
04:19
<&ToxicFrog>
Clojure, Scala, and Hylang all target existing VMs not designed with tail call elimination in mind
04:21 VirusJTG [VirusJTG@Nightstar-09c31e7a.sta.comporium.net] has quit [[NS] Quit: Program Shutting down]
04:59 Derakon is now known as Derakon[AFK]
05:12 ktemkin is now known as ktemkin[awol]
05:18 Kindamoody[zZz] is now known as Kindamoody
05:20 ErikMesoy|sleep [Erik@A08927.B4421D.FE7332.0AD079] has quit [Ping timeout: 121 seconds]
05:26 ErikMesoy [Erik@Nightstar-404c8d4f.80-203-17.nextgentel.com] has joined #code
05:27 ErikMesoy [Erik@Nightstar-404c8d4f.80-203-17.nextgentel.com] has quit [Client closed the connection]
05:38 Turaiel [Brandon@Nightstar-949d7402.resnet.mtu.edu] has quit [[NS] Quit: Bouncer terminated]
05:40 Turaiel [Brandon@Nightstar-949d7402.resnet.mtu.edu] has joined #code
06:12 Turaiel is now known as Turaiel[Offline]
06:23 KiMo|autorejoin [Kindamoody@Nightstar-05577424.tbcn.telia.com] has joined #code
06:24 Kindamoody [Kindamoody@Nightstar-05577424.tbcn.telia.com] has quit [Ping timeout: 121 seconds]
06:25 KiMo|autorejoin is now known as Kindamoody
08:25 Karono [Karono@Nightstar-13c26ed9.optusnet.com.au] has joined #code
08:34 den [NSwebIRC@Nightstar-9fdec4be.frokca.sbcglobal.net] has joined #code
08:35 den [NSwebIRC@Nightstar-9fdec4be.frokca.sbcglobal.net] has left #code [""]
09:01 You're now known as TheWatcher
09:14 mode/#code [+o Kindamoody] by ChanServ
09:48 Kindamoody is now known as Kindamoody|out
10:18 AverageJoe [evil1@Nightstar-4b668a07.ph.cox.net] has joined #code
11:11 AverageJoe [evil1@Nightstar-4b668a07.ph.cox.net] has quit [[NS] Quit: Leaving]
12:44 Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving]
12:59 celticminstrel [celticminst@Nightstar-ae361035.dsl.bell.ca] has joined #code
12:59 mode/#code [+o celticminstrel] by ChanServ
13:34 Karono is now known as MoonMoon
13:36 MoonMoon is now known as Karono
15:14
< AnnoDomini>
Why does imgur uploading take more time per image when you upload a whole lot of them?
15:28
<@gnolam>
Is it sending them in parallel?
15:37
< AnnoDomini>
No, sequentially.
15:38
< Xon>
probably a deliberate back-off
16:33
< [R]>
"Because of encrypted connections zoosk coins hack is impossible to be detected."
16:33 Karono [Karono@Nightstar-13c26ed9.optusnet.com.au] has quit [Client closed the connection]
17:44 Turaiel[Offline] is now known as Turaiel
18:08 AnnoDomini is now known as Wires
18:59 Kindamoody|out is now known as Kindamoody
19:27 Kindamoody is now known as Kindamoody[zZz]
20:16 himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds]
20:29
<&ToxicFrog>
Aaaaaaaaaaaaaa
20:30
< Wires>
Bbbbbbbbbbbb
20:30
<&ToxicFrog>
Stuff is on fire and me and my immediate boss (who just transferred into this team last month) are the only people in the office
20:30 himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code
20:30 mode/#code [+o himi] by ChanServ
20:35 himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds]
20:47 himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code
20:48 mode/#code [+o himi] by ChanServ
21:11 himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds]
21:14
< Wires>
Hmm. I didn't get an answer yesterday... Is SDL2 worth upgrading to?
21:16 Turaiel is now known as Turaiel[Offline]
21:16 Turaiel[Offline] [Brandon@Nightstar-949d7402.resnet.mtu.edu] has quit [[NS] Quit: Bouncer terminated]
21:16
<&McMartin>
Probably
21:17 Turaiel[Offline] [Brandon@Nightstar-949d7402.resnet.mtu.edu] has joined #code
21:19
< Wires>
I'm guessing there's a "depends" there somewhere?
21:19
< Wires>
'Cause SDL1 seems to have everything I need.
21:20
< [R]>
Might be worth it to abstract it enough that swapping won't be a bitch?
21:34
<&McMartin>
That's actually pretty hard
21:34
<&McMartin>
So, I don't have a lot of time right now, but, briefly:
21:34
<&McMartin>
SDL1 supports a shitload of ancient platforms
21:34
<&McMartin>
SDL2 doesn't, but *does* have targets for iOS and Android native
21:34
<&McMartin>
SDL2 has two things that are compelling enough that I'll switch my stuff to it eventually
21:35
<&McMartin>
(1) Logical vs. Physical screen size, so I can program to, say, 480p as a pixel art game and it'll run 1080p fullscreen with no changes to render logic
21:35
<&McMartin>
(2) XInput support so that the LT and RT buttons on my gamepad work
21:35
<&McMartin>
But the translation isn't generally trivial; SDL2 explicitly divorces GPU-stored images from CPU-stored ones.
21:36
<&McMartin>
SDL1 tries to do it on its own and isn't great at it. This usually doesn't matter, but it can.
21:36
<&McMartin>
But even if you *do* need stuff in SDL2, it's not a great time to switch, because the support libraries aren't there yet.
21:37
<&McMartin>
Oh right
21:37
<&McMartin>
It's not compelling *for me*, but another possible compelling reason would ben
21:37
<&McMartin>
*be
21:37
<&McMartin>
(3) You want to use OpenGL 3.0 or later and use SDL to handle the windowing system bullshit
21:37
<&McMartin>
(4) You want more than one window open at once
21:38
<&McMartin>
(SDL 1 has a global notion of "the screen" and it also can only set up contexts for OpenGL 2.1 or earlier)
21:38
< Wires>
OK, so my answer would be no.
21:39
<&McMartin>
Yeah, reason (1) is enough to convince me to switch eventually, but I've built out SDL_mixer and SDL_image wrappers and stuff that wont' work with SDL2_mixer and SDL2_image, neither of which are in my repos, either.
21:41
<&ToxicFrog>
And for me (1) and (4) would be killer if I still used SDL for anything.
21:55
<&McMartin>
The real issue now is that SDL1 is noticably obsolete (and has been since, oh, Windows 7 came out, maybe a little before) and SDL2 is still bleeding edge
21:56
<&McMartin>
So "wait for SDL2 to stop bleeding" is my real plan.
21:56
< Wires>
Well, I still use WinXP. :P
21:57
<&McMartin>
Yeah, SDL1 evolved to match that. But it was designed for Win95 and Linux with OpenGL 1.x.
21:59 Wires is now known as AnnoDomini
22:14 Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code
22:14 mode/#code [+qo Vornicus Vornicus] by ChanServ
22:14
<@Tarinaky>
I don't think SDL2 is going to stop being bleeding edge.
22:15
<@Tarinaky>
Too much Second System Syndrome.
22:15
< AnnoDomini>
Has to, eventually.
22:15
< AnnoDomini>
It'll eventually gather enough crust to scab over.
22:15
< AnnoDomini>
They'll be rolling out SDL3 in 10-15 years. :P
22:15
<@Tarinaky>
It can (easily) fail first.
22:16
<@Tarinaky>
SDL3 doesn't require SDL2 to ever be in a stable state.
22:16
< AnnoDomini>
A corpse stops bleeding.
22:18
<@Tarinaky>
Doesn't make it useful.
22:18
< AnnoDomini>
Does make it non-bleeding edge.
22:19
<@Tarinaky>
If all you want is for a non-moving target you can fork a revision (any revision) and not update it.
22:19
<&McMartin>
Tarinaky: How do you figure for second system syndrome?
22:19
<@celticminstrel>
...is there an SDL3?
22:19
<&McMartin>
Most of the complaints I've seen are that not *enough* changed.
22:19
<&McMartin>
SDL2 literally came out a couple weeks ago, there is not an SDL3 right now.
22:20
<@celticminstrel>
Oh huh, it's finally actually released?
22:20
<@Tarinaky>
Oh, I didn't know that.
22:21
<@Tarinaky>
The original SDL was borne from neccesity and a very specific use case...
22:21
<@Tarinaky>
SDL2... not so much.
22:21
<&McMartin>
Heh. That changed when Valve hired Sam Latinga~
22:21
<&McMartin>
SDL2 suddenly got one, very specific use case: TF2
22:21
<@Tarinaky>
Ahh.
22:22
<@Tarinaky>
This makes mountains of sense!
22:22
<@Tamber>
Suddenly, all the hats become clear.
22:22
<&McMartin>
From what I've seen of the API, most of the changes are splitting out SDL_Surface into SDL_Surface, SDL_Texture, and SDL_Renderer.
22:23
<&McMartin>
Which is fine, imo; SDL1's SDL_Surface abstraction was just a little bit too high to get good consistent experiences across the devices I used it with.
22:24
< AnnoDomini>
In slightly related news: I've managed to get Vinland Trail to compile for Windows. If anyone's interested, I can send a link.
22:24
<&McMartin>
(Interestingly, while TF2 runs fine on my Linux machine with an Intel HD4000, Portal glitches out a lot)
22:26
< Syka>
McMartin: are you sure those glitches arent actually portal arg: electric boogaloo
22:26
< Syka>
have you checked the coredumps for BBS logins
22:26
< Syka>
portal arg 2*
22:26
<&McMartin>
Hee
22:26
<&McMartin>
Pretty sure.
22:26
<&McMartin>
I suspect it's the drivers messing up render-to-texture
22:27
<&McMartin>
Trying to manipulate objects on the other side of a portal gap makes the rendering flicker
22:32 * Vornicus makes the first of several posts about solving klotski do his blogthing
22:34
<~Vornicus>
http://dantheunspeakable.com/blog/programming/klotski-1--introduction--loading-- and-internal-representation/
22:40
<&McMartin>
A thing that came up in my LISP experiements; if you don't do that shape-based collapse it actually becomes computationally infeasible.
22:46
<~Vornicus>
576 fold improvement is pretty major, yes.
--- Log closed Sat Aug 31 00:00:10 2013
code logs -> 2013 -> Fri, 30 Aug 2013< code.20130829.log - code.20130831.log >

[ Latest log file ]