Question about the JME3 sceneGraph

Hi guys,

I wrote a jogl program by myself and I implemented a very basic sceneGraph.

Looking for improvements, I read in the past days a couple of slide about advanced rendering pipeline, especially this one.

I found it very interesting and I was wondering how much advanced is the sceneGraph of JME3, in order to decide if jump to it or write it by myself…

I found only the wiki page, that gives just a basic explanation about what is a sceneGraph and the transformation order, but I would like to know if there is anything interesting for me under the hood…

You can always read the source and see if you find whatever it is that is interesting to you.

1 Like

For me, this sounds as “Convince me to use jME3”. To be honest, I had no idea what jME3 was capable of doing when I jumped into it.

What make me stay on jME3? Two key points:

1 - Source Code available, so if I don’t understand something, just dig deep into it;
2 - Community: We are awesome!

So what I can say for you is give it a try, you won’t regret.


Whether or not there is interesting stuff under the hood I guess would depend on what interests you. It is a very tight scene graph implementation as they go.

Rendering, data updates, picking, etc. are done without the visitor pattern so it avoids double dispatch in favor of classic (and efficient) polymorphism. These things also do array iteration to avoid garbage creation. Recent improvements even tighten up how logic updates are done and avoid data structure updates in more cases where it’s unnecessary.

But really, it’s easy to try out, I guess. See if it’s better than the basic one you hand-rolled.

1 Like