What are you working on? January 2012

[thelinx]

[QUOTE=ZenX2;34095953]Okay, I'm switching from love's filesystem to my file library for configuration/other type files.[/QUOTE]

Why?

[zenby]

[QUOTE=thelinx;34095985]Why?[/QUOTE]

It makes the matter of dealing with having 2 copies of all of my configuration files. I keep them in the .love, and writing to them creates new ones in AppData. Also, I prefer having things like save files where the game is installed. Having saves/etc. in AppData makes it more difficult for users who have might have trouble finding them, also love forces you to keep everything inside the LOVE app data folder.

[r0b0tsquid]

[QUOTE=Jawalt;34095596]Well I wrote a grapher a while back, and when I had the resolution of the graphs 1:1 with pixels it was GODAWFUL slow, even when offloaded to the GPU. (On a 320M, which is slightly above what's in the average computer). So I could see if you were to just check for common graphs with regex or something and instead of solving for each horizontal row of pixels you can just draw a line.[/QUOTE]
Rated artistic, because I'm not sure whether that's a good idea or really stupid :v:

[esalaka]

[QUOTE=r0b0tsquid;34096955]Rated artistic, because I'm not sure whether that's a good idea or really stupid :v:[/QUOTE]

If an equation is of the form [xy]=<statement>, you only have to process it for all the points on one axis. Checking common graphs with a regex would be silly but recognizing where implicits are not used is actually a good idea performance-wise.

[Darwin226]

[QUOTE=Jawalt;34095596]Well I wrote a grapher a while back, and when I had the resolution of the graphs 1:1 with pixels it was GODAWFUL slow, even when offloaded to the GPU. (On a 320M, which is slightly above what's in the average computer). So I could see if you were to just check for common graphs with regex or something and instead of solving for each horizontal row of pixels you can just draw a line.[/QUOTE]

What exactly is the point of having a grapher that's awfully slow unless you input a common formula?

[Nexus435]

[QUOTE=COBRAa;34076358]Yes, I understand this perfectly, I just wasn't aware the client set his own seed, I thought it was assigned by the server and the server uses the same one it sent to the client to do the spread.[/QUOTE]

But the effect would be clientside if you set it to the same number?

[r0b0tsquid]

It already does that :) not with a regex though :v: I added explicit plots first, because they're just a lot simpler than implicit ones. The only issue is that if you have a relation y = f(y) or something, plotting it as an explicit gives interesting (read: incorrect) results, as someone pointed out a couple pages ago.

[editline]7th January 2012[/editline]

Also, all of the 3D plots you've seen from it so far have been explicit, because marching cubes is enormous (256 separate cases :S)

[Gran PC]

Hey guys, I need feedback on the interface for the Facepunch monitor:

[img]http://www.peniscorp.com/stuff/ZScreen-2012-01-07_21.28.29.png[/img]

Does this look good?

[Overv]

[QUOTE=Jawalt;34095596]Well I wrote a grapher a while back, and when I had the resolution of the graphs 1:1 with pixels it was GODAWFUL slow, even when offloaded to the GPU. (On a 320M, which is slightly above what's in the average computer). So I could see if you were to just check for common graphs with regex or something and instead of solving for each horizontal row of pixels you can just draw a line.[/QUOTE]

You were doing some very strange things then, because I can easily plot explicit expressions with GLSL:

[url]http://glsl.heroku.com/e#1007.0[/url] (Set resolution to 1)

Sure, it isn't perfect yet, but it's fast and it works. [url=http://glsl.heroku.com/e#1007.1]Here[/url] is a hand calibrated version, although that's cheating a bit.

[r0b0tsquid]

Using Newton-Raphson to find the closest intersection of two curves to the mouse position:
[img]http://i.imgur.com/gcbGd.png[/img]
(tiny pink dot is the mouse position)
Also added a grid to the 3D plots, working on time atm :)

[LuaGuy]

[QUOTE=GranPC;34097791]Hey guys, I need feedback on the interface for the Facepunch monitor:

[img]http://www.peniscorp.com/stuff/ZScreen-2012-01-07_21.28.29.png[/img]

Does this look good?[/QUOTE]

yeah except the picture

[Darwin226]

[QUOTE=Overv;34098443]You were doing some very strange things then, because I can easily plot explicit expressions with GLSL:

[url]http://glsl.heroku.com/e#1007.0[/url] (Set resolution to 1)

Sure, it isn't perfect yet, but it's fast and it works. [url=http://glsl.heroku.com/e#1007.1]Here[/url] is a hand calibrated version, although that's cheating a bit.[/QUOTE]

Added AA to and more precise graphs.
[url]http://glsl.heroku.com/e#1009.0[/url]

[editline]7th January 2012[/editline]

Not really AA.

[ThePuska]

[QUOTE=r0b0tsquid;34098714]Using Newton-Raphson to find the closest intersection of two curves to the mouse position:
[img]http://i.imgur.com/gcbGd.png[/img]
(tiny pink dot is the mouse position)
Also added a grid to the 3D plots, working on time atm :)[/QUOTE]

When you mentioned the Newton-Raphson method, I thought of another way to plot implicit equations. I'm not sure how well - if at all - it would work.

One could calculate the partial derivatives of the implicit equation with respect to each graph variable and use the Newton-Raphson method to converge towards some point, or some points, that fulfill the equation.

After that one could calculate the total derivatives of the graph variables. Given a point on the graph, and the total derivatives, you have a tangent - which you can use to take short step that should not diverge from the graph very much. Iterate the Newton-Raphson method a few times and you're back on the graph, but hopefully having travelled forwards along it.

I don't have a grapher to test this on :C

[Rohan]

[QUOTE=swift and shift;34094261]<3 php

[url]https://bugs.php.net/bug.php?id=60682[/url][/QUOTE]
nothing new, php was always shit

[Murk]

[QUOTE=ZenX2;34096212]It makes the matter of dealing with having 2 copies of all of my configuration files. I keep them in the .love, and writing to them creates new ones in AppData. Also, I prefer having things like save files where the game is installed. Having saves/etc. in AppData makes it more difficult for users who have might have trouble finding them, also love forces you to keep everything inside the LOVE app data folder.[/QUOTE]

Don't do anything stupid. The reason Appdata is used is so that user data can easily be backed up since all programs are [I]supposed[/I] to be saving there. Granted, there are loads that don't, but that doesn't mean you should as well. Just generate the configuration files on first run if they don't exist. That way you can immediately account for user's resolution and such.

If it's user's application data it should be saved in C:\Users\username\Appdata. That's why it's there for. If users are confused, tell them to open up file manager and write "%appdata%".

If all games and programs saved their data (obviously not all data, but user related data) that way, Windows would be a better place.

[r0b0tsquid]

[QUOTE=ThePuska;34099022]When you mentioned the Newton-Raphson method, I thought of another way to plot implicit equations. I'm not sure how well - if at all - it would work.

One could calculate the partial derivatives of the implicit equation with respect to each graph variable and use the Newton-Raphson method to converge towards some point, or some points, that fulfill the equation.

After that one could calculate the total derivatives of the graph variables. Given a point on the graph, and the total derivatives, you have a tangent - which you can use to take short step that should not diverge from the graph very much. Iterate the Newton-Raphson method a few times and you're back on the graph, but hopefully having travelled forwards along it.

I don't have a grapher to test this on :C[/QUOTE]
[editline]6th January 2012[/editline]
So you'd iterate until you found an edge, then "walk" your way around it by using Newton-Raphson to glue yourself to the edge? The main problems I see are (a) knowing when you've found all of possibly multiple edges and (b) knowing when you've found the same edge more than once. To find all/most of the edges, you'd have to sample a grid of points anyway, and differentiation requires that the expression be evaluated twice, so the current method is probably more efficient. Your method would be much more accurate for fields that don't vary linearly, though, and I really like the idea :)

[acpm]

[QUOTE=GranPC;34097791]Hey guys, I need feedback on the interface for the Facepunch monitor:

[img]http://www.peniscorp.com/stuff/ZScreen-2012-01-07_21.28.29.png[/img]

Does this look good?[/QUOTE]

I know you wanted me to do the rest of the UI for that project but I'll probably just be exposing IClientFriends to lopensteamworks today, so I'll review with you how that needs to change.

[Gran PC]

[QUOTE=LuaGuy;34098852]yeah except the picture[/QUOTE]

Thanks for your concern on my avatar. For criticism regarding my avatar, I can be reached at /dev/null. Cheers!

[pollyzoid]

Running Lua in TF2.

[img]http://puu.sh/cIJC[/img]

[lua]include "core/util.lua"

local hook = require 'hook'

print( "In menu:", IN_MENU )
print( "In game:", IN_GAME )
print( "Game:", GAME )

hook.add( "LevelInitPreEntity", "tset", function( map )
print( "LAU MAP", map )
end )

local concommand = require 'concommand'
concommand.add( "ir_test", function( cmd, argv, args )
print( "HI", args )
end, "Hello this is help.", function( partial )
return { "DOG", "GOD" }
end )

print.warning["test"]( "Weeeeeeeeeeeeeeeeeeeee", "*SPLAT*" )
[/lua]

So exciting :v:

[esalaka]

how did you lau in TF2

[jalb]

[img]http://dl.dropbox.com/u/11093974/Junk/grass3.png[/img]
The grass is actually temporary. Which is sad because my artist's five minute job is 100 times better than anything I can do...

[acpm]

[QUOTE=esalaka;34100169]how did you lau in TF2[/QUOTE]

Signature scanning and loading Lua just like any other module.

Feel free to use my signatures raBBish:
[url]http://hl2sb-src.googlecode.com/svn/trunk/src/utils/sourcelua/[/url]
[url]http://hl2sb-src.googlecode.com/svn/trunk/src/utils/sourcelua_swarm/[/url]

[IMG]http://i.imgur.com/7lugE.jpg[/IMG]
[IMG]http://i.imgur.com/qvmvG.png[/IMG]
[IMG]http://i.imgur.com/vAwTy.png[/IMG]

[Exl]

[img]http://www.dvdflick.net/storage/gb/gb_01.jpg[/img][img]http://www.dvdflick.net/storage/gb/gb_02.jpg[/img][img]http://www.dvdflick.net/storage/gb/gb_03.jpg[/img]

Yet another gameboy emulator + rudimentary debugger. It's a lot of fun figuring out how a gameboy works internally, and how emulators work under the hood. The gameboy is a simple enough device for it not to be months of work. So far the most work was making the giant opcode tables for ~500 opcodes and writing the functions for each one. That and debugging why a game doesn't work. So many hours of staring at Z80 code... It's now capable of playing a couple of games, though it does not yet display sprites\windows or play sounds.

[acpm]

-snip, added above-

[editline]7th January 2012[/editline]

[QUOTE=Exl;34100773][img]http://www.dvdflick.net/storage/gb/gb_01.jpg[/img][img]http://www.dvdflick.net/storage/gb/gb_02.jpg[/img][img]http://www.dvdflick.net/storage/gb/gb_03.jpg[/img]

Yet another gameboy emulator + rudimentary debugger. It's a lot of fun figuring out how a gameboy works internally, and how emulators work under the hood. The gameboy is a simple enough device for it not to be months of work. So far the most work was making the giant opcode tables for ~500 opcodes and writing the functions for each one. That and debugging why a game doesn't work. So many hours of staring at Z80 code... It's now capable of playing a couple of games, though it does not yet display sprites\windows or play sounds.[/QUOTE]

That is mighty impressive.

[Jawalt]

[QUOTE=Overv;34098443]You were doing some very strange things then, because I can easily plot explicit expressions with GLSL:

[url]http://glsl.heroku.com/e#1007.0[/url] (Set resolution to 1)

Sure, it isn't perfect yet, but it's fast and it works. [url=http://glsl.heroku.com/e#1007.1]Here[/url] is a hand calibrated version, although that's cheating a bit.[/QUOTE]

Well, what I was doing (and only got to work in a very simple limited manner) was taking user input, RPNing it, and then 'compiling' it to a very simple shader. What I was doing that was so slow, I realize now and hadn't even thought about at the time was that I was drawing a fullscreen quad for each graph :V.

[zenby]

[QUOTE=Murkrow;34099374]Don't do anything stupid. The reason Appdata is used is so that user data can easily be backed up since all programs are [I]supposed[/I] to be saving there. Granted, there are loads that don't, but that doesn't mean you should as well. Just generate the configuration files on first run if they don't exist. That way you can immediately account for user's resolution and such.

If it's user's application data it should be saved in C:\Users\username\Appdata. That's why it's there for. If users are confused, tell them to open up file manager and write "%appdata%".

If all games and programs saved their data (obviously not all data, but user related data) that way, Windows would be a better place.[/QUOTE]

Okay. It seems every time I tell you guys a decision I've made it turns out that I'm doing it backwards :v:

[Tezzanator92]

Obviously the solution is to make a decision and do the exact opposite, then :v:

[zenby]

I'll try that.

Also, is there any way to have an AppData folder outside of the LOVE folder? It feels really unprofessional.

[editline]7th January 2012[/editline]

Oh yeah, I got a year of Microsoft Dreamspark. Thanks FIRST <3

[Murk]

LOVE 0.8.0 will put your folder directly in Appdata.

[acpm]

[QUOTE=ZenX2;34102275]I'll try that.

Also, is there any way to have an AppData folder outside of the LOVE folder? It feels really unprofessional.

[editline]7th January 2012[/editline]

Oh yeah, I got a year of Microsoft Dreamspark. Thanks FIRST <3[/QUOTE]

There is a release mode for this purpose. If I remember correctly, it takes away things like this; makes your folder its own AppData directory, and not a subdirectory in LOVE.

[editline]7th January 2012[/editline]

In 0.8.0, I guess you don't have to worry about that.