Display system "object model"

I am having a hard time trying to understand the "object model" for DisplaySystem – who contains whom, and what their multiplicities are.  It seems DisplaySystem is a singleton and, as such, Renderer and its Camera are effectively "global" references that would be used by all views (canvases).  As such, it seems, that a given application can potentially have multiple "views" of a single 3D "world", but this is conjecture,  especially since I am using SWT, which does not seem to use DisplaySystem.createWindow(), and instead uses a canvas impl approach, which is essentially undocumented in the wiki.

I have come across references to such things as RenderContext and JMEContext, as well as OpenGL contexts, but it is unclear how all of this is meant to fit together, or whether it is all meant to work together.

In summary, is there a straightforward way to achieve the following in JME:

  1) One 3D "world" with multiple "views".

  2) Multiple 3D "worlds", each with its own independent "view".

  3) Multiple 3D "worlds", each with multiple independent "views".

I have the same question (


If that is impossible (Multiple 3D "worlds", each with its own independent "view"), I am stuck, and I must choose another technology :frowning:

It currently does not but it has been a feature that is greatly desired, and although delayed because of the recent 'change in management', it is still on the table…

I guess you’re looking for something like this?

It’s possible with JMEContext, unfortunately an official distrubtion of the jME2 version was delayed… The only way to get it is by checking out the radakan source and getting the JMEContext project:




okay, now its REALLY time to figure out what to do on the old stuff; so we can move on…

(I had no idea you had already solved this momoko_fan, NICE WORK!)