The real kicker is that emscripten supports SDL by default, which means that most of the display and hopefully even keyboard stuff will be transparent.
There are a number of obstacles that will need to be overcome. I'll make a list of a few here:
- Will PEBL really compile and run, with its use of the many complex libraries, templates, flex/bison generated code, a fairly low-level counted-pointer reference system, hardware interactions, etc.?
- How can we make PEBL asynchronous? The recursive evaluator which steps through a PEBL experiment provides no easy way to cede control to the browser. This will likely be the obstacle that prevents success.
- How can we easily support running/selecting experiments for users