Not nearly as impressive as most of the other stuff in this thread, but I just finished the realtime lighting engine for my stealth game! Whoo! (Linked for ~1MB gif) https://wi.somethingawful.com/17/1782c8630c19e60efb5ab8a877136696a76aba73.gif EDIT: With a different tileset: Jo fucked around with this message at 21:27 on Jun 28, 2008 |
|
# ¿ Jun 28, 2008 19:59 |
|
|
# ¿ May 6, 2024 01:44 |
Image search engines are . You guys will likely get another 'break/rape/pwn my webapp' thread in the next month or so. http://www.omgitswebsite.com/HotSauce/
|
|
# ¿ Jan 20, 2009 01:43 |
necrobobsledder posted:Real men challenge other CoCers to do something and hope someone gets e-hurt enough that they write it out of pride. Real men are a superset of rational men. Rational men are a superset of natural men.
|
|
# ¿ Jan 27, 2009 04:02 |
Roflex posted:Going back to working on Generations. Good idea. Impressive execution. (And good choice of music for the video.)
|
|
# ¿ Jan 29, 2009 02:02 |
brian posted:Small point of contention: The shadows are falling in two different directions. Consistency here can improve the (already high) quality of your presentation.
|
|
# ¿ Mar 23, 2009 21:46 |
heeen posted:Extending on my previous work, now refraction + caustics: That's very much a thing of beauty. Didn't see much in the way of documentation in /proj, though. Have you a source or publications page?
|
|
# ¿ Apr 20, 2009 01:48 |
Not nearly as impressive as a lot of the stuff here, but I recently got moved to the team doing object recognition and reconstruction. Only user input is a single click to select what object to recognize. With simpler objects like chairs and tables, the application can reconstruct 3D models using a single image. It's still very much in the beta.
|
|
# ¿ May 25, 2009 22:19 |
When it comes to arguing with AD, the only winning move is not to play.
|
|
# ¿ May 29, 2009 04:30 |
Better to flatter someone who doesn't need it than to make fun of someone.
|
|
# ¿ May 29, 2009 23:51 |
Screeb posted:What's the hue if there are multiple routes to the same location? Or does it only work for mazes with one path to each location? I'm guessing it uses the lowest distance to the location, unless he's using something other than A*.
|
|
# ¿ Aug 29, 2009 17:54 |
Scaevolus posted:I'm using Dijkstra's algorithm because it lets you find the shortest path between a vertex and every other vertex on the graph, while A* just uses heuristics to find a path between two vertices on a graph. A* is shortest path, too. It's also faster than Dijkstra, if I remember right. Jo fucked around with this message at 15:21 on Aug 30, 2009 |
|
# ¿ Aug 30, 2009 15:19 |
darqness posted:I have done a lot more work on my maze game including MULTIPLAYER and an editor that is way faster. Also create an account to customize your guy or you can just stay anonymous as a blue square. The starting area is pretty turbulent so if you want to draw something cool just head to the west or east for some clean area. WASD moves and Space draws. I made something in the screenshot!
|
|
# ¿ Dec 5, 2009 23:20 |
Contains Acetone posted:
I've been trying to get a hold on Hinton's work since I saw his Google talk. I might pester you for code when you're done.
|
|
# ¿ Jan 21, 2010 04:24 |
Finishing up a sprite animation editor which imports/exports XML. Tile games are fun. EDIT: Added Internet Janitor's suggestions. Have origins now. No auto-bounding box yet. Also cropped for tables and added blue blobs for added work safe-ness. Jo fucked around with this message at 22:52 on Oct 23, 2010 |
|
# ¿ Oct 23, 2010 05:02 |
Internet Janitor posted:Jo: On the one hand, that seems reasonably useful. On the other hand, you created something new that makes XML and will inevitably create other new things that read XML. TilED, the map editor, exports XML. I figured rather than hack together some binary format it would be easier to just re-use the code. Beside, Python has an easy XML implementation and Java can just copy/paste my loading code. As for the origin, you're right. Thanks for the suggestion. Jo fucked around with this message at 18:13 on Oct 23, 2010 |
|
# ¿ Oct 23, 2010 17:59 |
Wrote an iterative permutation generator (rather than using itertools) to cycle through an alphabet. Added support classes and hosed with parallel python to create a distributed hash cracker. Whoo lunch break! It's running locally here because I'm a lazy toss who doesn't want to pull out his laptop.
|
|
# ¿ Mar 12, 2011 04:09 |
I'm peeved that DNF and Deus Ex: Human Revolution are so far away. I decided to start (or restart, depending on how much code I use from Devil Gear Solid) and make a 2D stealth game. Hoping to have a decent demo by the end of the weekend. I suck at drawing sprites. Jo fucked around with this message at 08:07 on Apr 16, 2011 |
|
# ¿ Apr 2, 2011 06:52 |
Seven Round Things posted:Any progress on this? I've always been bummed out by the lack of indie stealth games. Yes. I'd say I'm close to 35-40% done. The basic pieces are in place, but I need to refactor the mapping system and make the entities less stupid. EDIT: I should say I'm aiming for release of a one level demo on May 15. Jo fucked around with this message at 23:52 on Apr 9, 2011 |
|
# ¿ Apr 9, 2011 22:53 |
wacky posted:This might be a dumb question but I can't tell if that is a real city or not. What city is it? Cityscapes are so effing cool and they tie in perfectly to the cyberpunk (dream) lifestyle of the programmer I have no idea. Google image search for Future Cityscape, or something like that. It, along with the title, "Sneaking Mission" are placeholders.
|
|
# ¿ Apr 10, 2011 01:50 |
Sneaking Mission posted:You should keep it. You make a convincing argument.
|
|
# ¿ Apr 10, 2011 16:37 |
Your Computer posted:
I see you, too, have discovered the absolute joy of the Slick 2D library. Rock on Slick buddy.
|
|
# ¿ Apr 13, 2011 08:11 |
Cowcatcher posted:Is there something similar to Slick 2D only in C? Check out wikipedia's list: http://en.wikipedia.org/wiki/List_of_game_engines
|
|
# ¿ Apr 14, 2011 04:23 |
Finally made my tileset editor usable. I love sprite games and there are a bunch of good options for map editors (like Tiled), but very few for tileset editors. Pixothello was awesome, but it ran like poo poo on my machine and couldn't import/export png.
|
|
# ¿ Dec 8, 2011 09:05 |
Contains Acetone posted:Working on a user-friendly, GPU-Accelerated, Restricted Boltzmann Machine training system that follows the guide put out by Dr. Hinton: http://www.cs.toronto.edu/~hinton/absps/guideTR.pdf This is awesome. I've tried (and failed) repeatedly to do a functional implementation. Any chance you'll put your code online?
|
|
# ¿ Dec 13, 2011 06:30 |
EDIT: :facepalm: http://cgcad.thss.tsinghua.edu.cn/chenjt/isppm/preprint.pdf
|
|
# ¿ Mar 14, 2012 07:42 |
From Earth posted:Preeeety.
|
|
# ¿ Mar 18, 2012 19:16 |
Some images from an explosion generator. Posting about it on #gamedevgoons.
|
|
# ¿ Apr 10, 2012 00:09 |
Why not just use the OpenGL orthographic camera? A perspective camera at optical infinity is practically the same thing, but you need to mess with floating point imprecision.
|
|
# ¿ May 8, 2012 16:29 |
Roflex posted:Maybe I'm misunderstanding the Ortho view but the way I did it, it's still 3d/perspective at the core, allowing free view from any angle. I'm not one to comment on your implementation -- I haven't seen it. My curiosity is why would one bother doing strange tricks to make a perspective camera into an orthographic one? Why not use an orthographic one in the first place? I assert the following:
Water under the bridge, though, since the implementation is done and fun was had doing it.
|
|
# ¿ May 9, 2012 02:16 |
Contains Acetone posted:
Oh god yes please.
|
|
# ¿ May 9, 2012 19:36 |
mortarr posted:
As neat as this is, do you seriously have a 20GB bandwidth cap? On a more related note, I've been working on a 4Chan/Reddit hybrid. Last night I got image search working. There are two algorithms in there right now, but it's a cakewalk to add more.
|
|
# ¿ May 17, 2012 18:15 |
I posted a link to a prototype game a while back. Ported it to Android and released on the Google Play Store. It's very basic, but it's kind of a fun time waster. There are two game modes: reflex and cone hunt. In reflex, you weave between cars on either side. In cone hunt, you try and smoosh construction cones.
|
|
# ¿ Aug 23, 2012 15:20 |
EDIT: Whoops. Meant this to be a PM.
|
|
# ¿ Sep 6, 2012 03:20 |
Perhaps one can cancel the delay on user input? Even staggering the messages by half a second would be almost imprescriptible in terms of gameplay, but would make it much easier to read.
|
|
# ¿ May 16, 2013 14:15 |
I've been working on a stupid puzzle game for Android and the iPhone. It's conceptually pretty simple. Take turns. Black moves first. Put down a piece, then slide a row or column. Pieces wrap around the edges. Get four in a row to win. There's also a super-secret beta copy if you've got Java: https://dl.dropboxusercontent.com/u/51917/ofu.jar Thank you to all the goons on IRC who have been playtesting. You guys are awesome.
|
|
# ¿ Jun 20, 2013 03:43 |
I've been playing with machine learning. I made a Twitter bot which scrapes NASDAQ and dumps a bad guess.
|
|
# ¿ Sep 5, 2013 04:38 |
unixbeard posted:This is cool and if you have any details of what you are doing behind the scenes I would be interested to hear them. You're going to be disappointed. It's a stupid simple thing. It starts by pulling stock price data, then for each company randomly samples tweets with relevant words, converts them to ngrams (order two), and attempts linear regression with the term frequency as the input matrix. For foreign terms (most of them, given the sparsity of the matrix), it pulls a number out of its rear end. The volume of the last trade is the 'confidence' until I stop being l lazy. unixbeard posted:To be on the safe side I'd make sure you don't put anything on there about anything you have a position in. I hadn't thought about that. Maybe I'll have to trade on NYSE instead.
|
|
# ¿ Sep 5, 2013 16:53 |
Taking a break from some of the harder math stuff to throw together a UI for labelling vehicle data. Summarily, the air force UAV dataset contains a ground truth, but it sucks dick. Average 'center' can be as far off as 100 pixels from the real vehicle. Since I'm not one to label 12,000 tracks by hand across 1TB of images, I made a backend that breaks things down into small chunks for processing by an army of undergraduates and the interface to add centers and forward vectors. There's still some usability stuff to be considered, like keyboard shortcuts and auto-switching from normal to create mode, but it works well enough for me to make a few hundred labels.
|
|
# ¿ May 28, 2014 05:39 |
I get really excited over stupid ideas and it takes me a long time to realize they're stupid.
|
|
# ¿ Oct 11, 2014 19:10 |
|
|
# ¿ May 6, 2024 01:44 |
Newf posted:Perplexing resume fodder. Does it paste the face over yours automatically or do you have to interact with it? It's automatic. Really just an excuse to play with the FaceDetection API.
|
|
# ¿ Oct 11, 2014 19:51 |