scene management

The current trend in scene management seems to be to use graphs to represent the current scene.So could someone enlighten me as to what makes a graph so much more useful than say a simple table?Or do I have it all wrong and graphs are only used for general-purpose scene management?
Say we have a bunch of scene nodes containing possition and orientation, geometry, shaders etc.We could easily stuff all these in a table and then simply traverse the table rendering all nodes,create a shader tree or whatever.So what do we gain by stuffing them in a graph instead?

If your wizard is carrying a staff and has some spell effects on himself, wouldn’t it be better to be able to cull the wizard first, and only worry about the staff/effects objects if their main node (the wizard) is actually going to be drawn?

Also, if you build a tree per portal in a portal engine, you don’t have to test all objects everytime you draw; only the objects for each portal zone you actually prove could be visible.

I see…btw is there any info on scene graphs on the net?So far I could only find the manuals of various scene graph apis to be of any use.


Try this link.

have fun.