Everything should be controllable with head-tracking. One of important features is text input, and we use dasher for that.
No inside action should cause lag or freeze.
By default, objects communicates by touching each other. This encourages visualization of internal objects.
Distant communication is done via JSON (similar to actor model) and not direct method call, so that no-freeze can be enforced.
Use minimum amount of code and data that feels good enough.
The code is licensed under FreeBSD.
(Especially) LibOVR has different license; see LibOVR/LICENSE.txt
See .travis.yml file for (ubuntu) package names.
Impl of Dasher (Eye-cursor only input method developed by some researchted). In construct, implemented using NativeScript + in-scene quad + very basic English language model.
