Re: [Gzz] Vobs and scene graphs

[B. Fallenstein]

Tuomas Lukka wrote:
> > Re the recent discussion on the jyu list about vobs and Berlin/Fresco 
> > (www.fresco.org). This made me think about the applicability of vobs to 
> > scene graph systems, again.
> >
> > In a sense a VobScene represents a scene graph. A vob system could be 
> > implemented in which each VobScene is basically a scene graph with keys 
> > from the model being attached at each node.
>
> Vobs need some more thinking: in the end it turned out that the key 
> paradigm made PP actually more *DIFFICULT* to implement; the key paradigm
> currently doesn't allow enough fine-tuning. For PP, we want to take more
> control, say "animate X to Y", i.e. the particular view that was clicked
> should be the one to animate to the center.

Yes, but wasn't that clear before? I mean, not even the simplest Zz 
views (repeating row/col) are implementable in the key paradigm 
painlessly. Or think about dimension lists and flobpaths.

What exactly do we need to do that the current system can't?

> > Another comment: I don't think Fresco supports connections between two 
> > different branches in the scene graph, from what I read. This is very 
> > important to us-- and could be interesting to them? But I would not know 
> > how to implement that in their architecture.

Actually, I see two ways. If there's a connection between B and C, both 
inside A, the connection could be a child of A and somehow try to find B 
and C inside A, or it could be a child of *both* B and C, but this would 
break the Fresco paradigm that if X is a child of Y and Z that's exactly 
the same thing as if there were two copies of X in Y and Z.

Or maybe, two objects inside B and C which have pointers to each other...

> True, this is important; vital to get performance tuned right for PP.

What does it have to do with performance?