| 16:19 | <annevk3> | I'm considering having breakfast as well... |
| 16:20 | <annevk3> | Though at this point I might just combine it with dinner. I was initially blaming this on the timeshift, but that obviously doesn't work anymore. |
| 16:24 | <eighty4> | annevk3: have brunch |
| 16:48 | gsnedders | wonders whether hitting 3000 words is still possible today |
| 16:49 | <Philip`> | Depends on how strictly you define "today" |
| 16:51 | <gsnedders> | Philip`: 2009-03-29T00:00:00Z/PT24H0M0S |
| 16:52 | <gsnedders> | (The definition will change at 2009-03-30T00:00:00Z |
| 16:52 | <jgraham> | It depends on how strictly you define words |
| 16:52 | <jgraham> | and hitting |
| 16:53 | Philip` | wonders if he's going to get horribly confused by his IRC client's clock being an hour away from local time |
| 16:54 | <gsnedders> | Philip`: Yes, you will. |
| 16:54 | Philip` | wonders how to change it |
| 16:55 | <jgraham> | My IRC client's clock is currently local time -7 |
| 16:55 | <Philip`> | "/script exec $ENV{'TZ'}='Europe/London'" - oh, that was easy |
| 16:55 | <fearphage> | Philip`: http://wiki.fishcracker.net/wiki/Irssi#Changing_the_timezone |
| 16:56 | <gsnedders> | Too easy, some might say. |
| 16:56 | <jgraham> | Oh, that worked |
| 16:56 | <Philip`> | fearphage: Hmm, that wiki seems to be wrong about saying I need to restart irssi |
| 16:56 | <jgraham> | Now I know when people said things |
| 16:56 | <jgraham> | Which is boring |
| 16:57 | <fearphage> | Philip`: no, you just need to change the timezone and export it |
| 16:57 | <fearphage> | i'm working on upgrading irssi without disconnection from anything |
| 16:59 | <fearphage> | any of you have ie >6 handy? |
| 16:59 | <Philip`> | jgraham: Try TZ=Asia/Kathmandu if you don't want to even know what minute somebody said things |
| 17:00 | <Philip`> | fearphage: When it says "to change the timezone displayed within irssi, you'll need to quit irssi, export TZ to the desired variable (and optionally add it to ~/.bashrc), and restart irssi" it does seem to be (incorrectly) telling me I need to restart irssi |
| 17:01 | <Philip`> | although maybe I'm misinterpreting what your "no" was in response to |
| 17:02 | <Philip`> | fearphage: Isn't /upgrade already there for upgrading without disconnecting? |
| 17:03 | <fearphage> | yea, i'm trying to coordinate with the guy that owns the server |
| 18:47 | gsnedders | shakes head at Brain Boyd's notes on Nabokov's Ada… |
| 18:47 | <gsnedders> | "Has masturbation ever before been rendered so artfully, so lyrically? And has the grand passion at the heart of a great love story ever been introduced so insistently in terms of masturbation?" |
| 19:49 | <jgraham> | Philip`: Does irssi do TZ=people/Hixie where say 08:00 would correspond to Hixie being at the point in his circadian cycle that would correspond to 8am for a normal person |
| 19:49 | <jgraham> | That would be highly useful |
| 19:52 | <Philip`> | jgraham: If you set up /usr/share/zoneinfo/people/Hixie then I'm sure it'd work fine |
| 19:55 | <jgraham> | Philip`: How well would that deal with a constantly changing offset from UTC? |
| 20:01 | <Philip`> | jgraham: The same as it deals with DST in other timezones, except more frequent |
| 20:03 | <Dashiva> | Does it handle multiple DST modes, though? |
| 20:04 | jgraham | assumes other dst systems are more or less periodic and don't need to be updated multiple times per day |
| 20:04 | <jgraham> | So it's not clear that a system that works for one use case will work for another |
| 20:05 | <jgraham> | s/an/the / |
| 20:06 | <annevk3> | I wonder if tying Web SQL to SQLite is bad whether tying immediate-mode Web 3D to OpenGL is bad too... Or does this comparison not hold? |
| 20:08 | <jgraham> | Arguably the opengl thing is worse since opengl sucks under windows |
| 20:09 | <annevk3> | That's my thinking. I wonder what roc thinks about this. |
| 20:09 | <Philip`> | OpenGL is a documented open standard with multiple implementations |
| 20:09 | <Philip`> | SQLite is just an implementation |
| 20:11 | <annevk3> | While that's certainly an argument, I'm not sure it matters much. |
| 20:12 | <Philip`> | It matters because it means you're not forced to use a specific implementation |
| 20:12 | <Philip`> | You could e.g. write something that provides an OpenGL-like API but implements it using DirectX |
| 20:12 | <Philip`> | and the OpenGL standard tells you how your implementation is meant to work, so you don't have to reverse-engineer someone else's |
| 20:13 | <annevk3> | That's the plan for Web SQL too... |
| 20:13 | <Philip`> | whereas the only way to implement something with an SQLite-compatible API is to reverse-engineer SQLite |
| 20:14 | <jgraham> | Philip`: That is a argument |
| 20:14 | <jgraham> | could you write an openGL wrapper with sensible perf though? |
| 20:14 | <Philip`> | I don't see why not |
| 20:14 | <Philip`> | Wine has a DirectX wrapper implemented using OpenGL, and that works pretty well |
| 20:15 | <Philip`> | (And the web stuff is OpenGL ES, which has much less of the cruft that normal GL has, so it's a closer match to DirectX) |
| 20:15 | <Philip`> | Of course shaders are likely to be a huge pain, though |
| 20:16 | jgraham | doesn't know much about this stuff |
| 20:16 | <jgraham> | Does anyone know of an atompub client that actually works? |
| 20:16 | <Philip`> | (since GL shaders are written in GLSL, and DX shaders are written in HLSL, which are different syntaxes for pretty much the same functionality) |
| 20:34 | <annevk3> | whoa, shader stuff looks complex |
| 20:34 | <annevk3> | what kind of language is that? |
| 20:35 | <annevk3> | (and is it guaranteed to be safe?) |
| 20:36 | <Hixie> | it's not |
| 20:36 | <annevk3> | Philip`, the problem is that writing wrappers for OpenGL is hard and likely to be ineffecient |
| 20:36 | <annevk3> | inefficient* |
| 20:36 | <Hixie> | browsers would need to precompile and verify them |
| 20:36 | gsnedders | is way too close to including the word "masturbate" in his English dissertation |
| 20:37 | <annevk3> | ouch |
| 20:38 | <annevk3> | can you actually verify it? |
| 20:40 | <Philip`> | annevk3: They're all C-like languages, though without stuff like pointers |
| 20:41 | <Philip`> | annevk3: I don't see why they'd be inefficient - the expensive parts of rendering are all handled by the hardware, and it doesn't matter whether the input came from DX-like API or a GL-like API |
| 20:41 | <annevk3> | Philip`, doesn't that mean it's impossible to figure out whether it will make things crash, leak memory, do evil things? there must be more to this since vlad & co know that won't work on the Web |
| 20:42 | <Philip`> | You can't leak memory because there's no memory allocation |
| 20:42 | <Philip`> | You can't crash because there's no memory protection |
| 20:42 | <annevk3> | Philip`, because you have to do more in software if the hardware doesn't have OpenGL drivers |
| 20:43 | <Philip`> | (Accessing arrays out of bounds is apparently undefined behaviour - I don't know what it does in practice) |
| 20:44 | <annevk3> | great |
| 20:44 | <Philip`> | You can do infinite loops but apparently at least some driver implementations have some kind of limit so you won't freeze your GPU |
| 20:46 | <annevk3> | I wonder at what point this Web thing becomes so complex it just falls over |
| 20:46 | <Philip`> | (kig (http://fhtr.blogspot.com/) has been doing some testing of GLSL-related stuff) |
| 20:47 | <Philip`> | annevk3: You wouldn't emulate the rendering in software - you'd translate everything into the native 3D-rendering API |
| 20:48 | <annevk3> | Philip`, since OpenGL is so low-level, that's hard |
| 20:49 | <Philip`> | The basic concepts are stuff like "here's a bunch of triangle vertex coordinates and a shader program that says what to calculate for each one, please draw them now", and every API is basically providing the same thing |
| 20:50 | <Philip`> | s/for each one/for each pixel/ |
| 20:53 | <Philip`> | I hope that one benefit of doing the 3d-canvas stuff in Khronos is that the driver developers (NVIDIA, ATI, Intel) might get involved and help produce a new shader language that can safely run untrusted code, and then it could be usable in a few years |
| 20:54 | <annevk3> | if we get them on board that'd be great, yeah |
| 20:56 | <takkaria> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=6746 is a complete headfuck |
| 21:19 | <annevk3> | takkaria, most of his feedback is, afaict |
| 22:32 | <annevk3> | http://www.zeldman.com/2009/03/04/dwws-3e/ will cover HTML5 apparently |
| 22:32 | <annevk3> | I guess that is the second book after the Danish one? |
| 22:35 | <annevk3> | I guess I'll buy a copy to see what it says. English is something I can read :) |
| 22:46 | <Niictar> | Heh, the Acid3 test makes the Wii version of Opera cry |
| 22:47 | <Niictar> | Or rather, running Acid3 in the Opera browsers makes the Wii cry and freeze and become unresponsive by test 40 |