Lately I spend some hours for a research on WebGL. It’s quite amazing in which ways internet technologies have evolved. I still remember times, when it was nearly impossible to render any graphics using JavaScript only. Nowadays it’s one of the minor difficulties since we have technologies like SVG or Canvas. Even after animated realtime graphics are nothing special anymore, there’s a new player in town: WebGL.
3d computer graphics in web applications seem to be still a topic thats is heavily associated with technologies like flash, shockwave or other plugin based platforms. This may become history soon. At least if the development of WebGL browsers will proceed as it currently does. If you already use a WebGL enabled browser and had a look at one of the several examples as they appear constantly on http://learningwebgl.com/, then you will figure out quickly what I’m talking about.
Beside the usual operations we are familiar with from nearly any 3D engine, the most Important part on WebGL seems to me that we are able to use vertex and fragment shaders in the usual native way. While the whole application code is written with JavaScript, the shader programs will be implemented in a C style shader code. In the end nothing special for OpenGL veterans. But hey! Exactly that’s the point! If you’ve ever developed your personal, most awesome shader ever, there’s the possibility to reuse or partly use them for WebGL. There’s an amazing experiment written by Inigo Quilez about this shader topic. It’s called Shadertoy (don’t forget to use a WebGL enabled browser). The Shadertoy website, currently offers you about 28 sample shaders of different kinds. All are looking very pretty and all are native shader programs, which are executed on the graphics accelerator. Hardware acceleration for Webapps: Just great!
The basic possibilities offered by this technology may end into enourmous impacts for webstandards. You already may have recognized the Flash Gordon Project of Tobias Schneider. His approach is to implement an open source flash runtime, written in JavaScript. Imagine this implementation, using the powers of GL rendering. Or ways better: a new runtime plattform using WebGL as rendering layer.
Enough of nerdy enthusiastic day dreaming. There are still some difficulties with WebGL. At first of course, the minor density of WebGL enabled browsers out there in userland. But there’s a silverline on the horizon: last September Mozilla announced to implement WebGL in future versions of Firefox. The current nightly builds yet implement it, but it’s disabled by default. So it’s just a matter of time, when WebGL will be available easily by default.
Second flaw: It is quite uncomfortable to keep track of all matrix transformations by yourself. Bare WebGL won’t do that job for you. Many examples on the internet (including the Learning WebGL lessons) implement their own helper functions for identity loading, by encapsuling Sylvester (vector and matrix math for JavaScript) routines. That’s a common but never the less hacky way to fix the problem. And even creating simple basic shapes still produces a lot of code. Many shaders are embedded directly into HTML code, like it is well know from CSS or JavaScript embedding. That feels pretty ugly, when you’ve a complex WebGL application in mind.
In my opinion, the best solution for this case is WebGLU by Benjamin DeLillo. This library is a set of frequently high usefull routines for developing your very own WebGL application. You can load externalized Shader code, enjoy the ease of a camera and many more fine features.
So if you are interested into WebGL now, you should have a look at Learning WebGL. It’s in my opinion one of the best and most up to date resources on the net. If you are interested in the papers and basics of the new standard, it’s the best to search The Khronos Group website for further information.
If you’ve any other interesting news or resources for WebGL, please leave a comment or message me on any of the known ways.
