DEKTEN

Lego Naming Prevents Higher Level Thinking : GunPowderedPoweredLargeIronBallLauncher <<-- Clean Code Cannon <<-- Terse Invented Concept In Your Code If you agree. Check out my coding stream. I do Data Oriented Game Engine Design. (DOGED) www.twitch.com/kanjicoder www.twitch.com/kanjicoder //:2021_02_23:===============================================:// WindowExternalVisibilityController ==> BLIND ExtraFlufflyCushionedChair ==> LazyBoy MultiBarrelRotatingMiniCanno ==> GatlingGun LargeIronBallLauncher ==> Trebuchet GunPowderedPoweredLargeIronBallLauncher ==> Cannon You cannot think of "GunPowderedPoweredLargeIronBallLauncher" at the next level of abstraction until you condense the word into a TERSE/SHORT word. Example: You won't invent a "Gauss Cannon" until you STOP CALLING YOU CANNON A: "GunPowderedPoweredLargeIronBallLaunched" Same applies to your code. ...

Too lazy to dig through all this? Visit me on twitch and I'll help you, since I obviously know! www.twitch.com/KanjiCoder www.twitch.com/KanjiCoder  I could get GL.texSubImage2D working... but when I tried to use the OpenGL version... I noticed an initial offset into the cpu-side bitmap (rgba byte array) was NOT a parameter to the OpenGL version of the function. Here are my fix notes for OpenGL call: QUESTION( via facebook): I can get WebGL's "texSubImage2D" to work... But no luck in OpenGL. The signature for WEBGL includes a "stride". But the OpenGL signature does not. Think that is what is messing me up? ANSWER( answered my own question later on facebook) Answer: Because OpenGL's version does NOT include an initial offset into the data like WebGL's version, you need to do some pointer math on the uint8_t byte array on CPU side that houses your image. I have a 512x512 bitmap on CPU side. I have a 512x512 texture on GPU side. To update a 32x32 sect...

 Auto tiles, also known as rule tiles, are pretty awesome. They speed up the game developement level design process at LEAST by a factor of 16x without ANY loss in quality of the final product. I am working on an engine that I hope will speed up the CREATION of auto tiles by 16x. I am obsessing over "data leverage". The idea that, if we use our information/data wisely we can speed up BOTH __HUMAN__ productivity and __COMPUTER_PROCESSOR__ productivity. I am done ranting. Here is the imgur post: https://imgur.com/gallery/29G0u6j Inlined as Image: Catch me on twitch: www.twitch.com/kanjicoder 

 Sometimes I see something with google image search, but then cannot find it again even though I will type the phrase in the comic verbatim. So... Putting this hear to see if google will index this. "Help Ethan, I am Stuck". https://imgur.com/gallery/1Ec19pw

Fractal Auto Tile Engine: I work on it almost every day on twitch: www.twitch.com/kanjicoder I am working on a metroid-vania style fractal auto tile engine. The basic idea is to draw the sub-tiles of an auto-tile set (auset) using other tiles instead of pixels. Before you can draw fractal auto tiles, you need to create terminal tile designs. This is where the PAINT5D system comes into play. It is responsible for holding 16 intricate auto-tile set (auset) designs. Each tile design has 3 layers per "hyperpixel" size. A "hyperpixel" being a pixel of a relative size rather than a fixed size. And which could be rendered as a tile instead of a pixel if need be.  There is still much work to do, but the diagram in the readme should help explain more. https://github.com/KanjiCoder/AAC2020/blob/master/README.MD For various technical demos used as proofs of concept I made before embarking on this [c99/c11] based engine, you can visit. https://d3m0.herokuapp.com/ Visit my twitc...

Easy. And the answer is so obvious to anyone skilled in the art that this should NOT be patentable . Take your auto tile sub tiles. Now instead of using different pixel colors for the auto tile sub-tile graphics, paint the graphic using OTHER AUTO TILES instead of pixels. I am recording this here because I don't want computer patent trolls fucking me over in the future. The easy part is the definition. The hard part is the execution. The whole moving from "first to file" to "first to invent" with the U.S. patent office seems... Kind of sketchy to me. Email Me: HeavyMetalCookies@Gmail.com

Working on a utility that can convert a .PNG file into a hex dump. You can then take that hex dump and #include it into an C99 array like such: uint32_t my_png_data[ ]={ #include "PNG_HEX_DUMP.TXT" }; I am doing this so I can embed assets into my game engine. My code it on github if you are interested in trying out my hex dump utility. The utility encodes as uint32_t rather than an array of bytes ( uint8_t ) in order to save memory used in the source code. 0xFF00FF00 is shorter than 0xFF,0x00,0xFF,0xFF GITHUB: github.com/KanjiCoder/PHEXDUM Twitch: www.twitch.com/kanjicoder

Around 2012 I was working on a game called "Atomic Alice" under the name "MakeGamesHappen" I still use some accounts with that name... But I've lost track of some passwords or been locked out of other accounts because of not logging in often enough. Crazy how you can know your password and security questions and still lose your account if you let it go in active. Google has kind of forced peoples hands. Forcing them to converge everything around GMAIL. You'll either do this because you are overwhelmed with the task of keeping track of accounts, or you'll keep track of your accounts, but lose them anyways because hey... use it or lose it. This is a rant. I don't feel like working right now. After flash died and the game project died... I wanted to find a tech stack to replace actionscript. I tried haxe, golang + sfml , golang + opengl , C++ , JavaScript + Three.js , JavaScript + Canvas , JavaScript + Webgl. A bit of gamemaker. Some unreal engine cod...