Friday, November 12, 2010

Now What?

For several months I've known that when I get home at night, I have tasks X and Y to do in Emscripten in order to move it forward. Last night, however, I took out the guitar instead. Suddenly, there is not much to do - basically the goals of the project have been achieved, Emscripten can compile things like the Bullet physics engine and run it on the web. As far as the core code-generating capabilities are concerned, Emscripten is pretty much complete.

So, what now?
  • There are some additional optimizations and enhancements that can be done, like nativizing structures or emulating multithreaded code.
  • There are various tooling improvements that can be done, like making it easier to glue together web code and compiled code.
  • Emscripten could be used in other ways, for example, it could be combined with something like Rubinius that generates LLVM code from Ruby, allowing running Ruby code on the web.
  • Various code cleanups and refactorings could be done.
I'm not in a rush to do any of these - none is urgent or essential. I guess I'll get around to them eventually, or perhaps someone else will.

Right now, I'm considering doing one or both of the following:
  • Return to my original goal, that of bringing Syntensity to the web. In other words, to compile a version of Syntensity using Emscripten. The time of 3D-environments-on-the-web is almost upon us, and when it is, we need to make sure that the main tools for it are open source and platform agnostic. Sadly, currently the main contenders are not such.
  • Some other side project, got at least two ideas in my head of things I'd like to hack up. They are very experimental and speculative though, so they may end up a waste of time. But if they succeed...

No comments:

Post a Comment