A brief hiatus from Ogoglio development has given me a bit of time to step back and evaluate how to drastically improve accessibility and ease the burden of development for people just getting into web based 3D spaces.
Step One: Usability
Over the next few weeks I'm going to frame out a Flash 10 based client which will replace the Java3D applet. The current client has played a crucial role in helping nail down exactly what we want on the client side (more on that later) but Flash's install base and recent betas make it clear that Flash is quickly becoming the de facto standard for 3D on the web.
Step Two: Refinement
Once we have a working Flash client, I have a new design for the server framework which will greatly simplify the code path between the DB and the browser. In short, the app tier will automatically wrap a specially crafted schema with a RESTful web api. On the client side, a Prototype based library will dynamically create a handy wrapper object. If my prototype is any indication, this new approach will not only reduce the load on Ogoglio developers but will be considerably easier on the app tier CPUs.
Step Three: Distribution
Once those pieces are in place I intend to combine the best of breed clustering platform for Amazon EC2 (currently Scalr is in the lead) with a precompiled Ogoglio system so that anyone with an EC2 account can boot a massively scalable node in Ogoglio City.
Comments