Hi all,

coming from life in the simplicity of the op stack or ice tree of soft I'm feeling decidedly uneasy about how it works in Houdini. A simple bit of modelling is clearly very easy, start at the top and work your way down, but how are things being ordered when you have multiple items going in and out of a dopnet? In a rigid body object it has two streams in it, the one generating the data that goes to dops and then the dops fetch data read back in but where is it being told to run this stream first? Also in a dopnet, why does gravity come after the solver? If someone knows how this all fits together I would be grateful for a heads up.


cheers,

A>

The DOP context is different (quite a bit actually), it describes behaviours and relationships so the order may or may not be important depending on the operator.

Furthermore, think of it as initial state and then a loop (that may run many time between frames) and how data out becomes the data in on the next loop.

I will suggest you “debug” what is going on by putting Wrangle nodes that count up an index so you can follow up how the tree is being evaluated.

A better analogy than the operator stack in Softimage would be Softimage simulation environment (in the explorer) because truly, before ICE, there was not such thing as a simulation stack right? (if my brain is still working I think this was the case)

there is this secret resource, no-one seems to know about these days, called Documentation

http://sidefx.com/docs/houdini/dyno/_index [sidefx.com]
http://sidefx.com/docs/houdini/dyno/top10_basics [sidefx.com]
http://sidefx.com/docs/houdini/dyno/top10_medium [sidefx.com]
http://sidefx.com/docs/houdini/dyno/top10_advanced [sidefx.com]
…

Thank you Tomas, much obliged for these. Normally I'm pretty good about RTFM but in this instance I thought perhaps some varied responses from soft users might add to a rounder understanding on top of what the manual has to offer.

and thank you Jordi, i like the idea of simple debugging to watch how things respond but I think I've a small way to go yet before I can do that.


cheers,

A

For those revisiting this I made a animation of what I believe the different systems are between SOPs and DOPs.
This was mostly made in reference to POPs and how DOPs handles geometry data, hope it helps

There is a fundamental difference in the way SOPs and DOPs handle geometry data:

In SOPs, the geometry starts at the top, and each subsequent node takes the geometry from its input(s), manipulates it, and passes it to the next node.

In DOPs, the geometry data is stored in the DOP object "geometry" data that is kept separate from the nodes. When the nodes are evaluated, they are processed from top to bottom, left to right. Each node, when evaluated, edits this geometry object.

This means that in DOPs, the nodes don't represent a flow of geometry passed from node to node, but rather an ordered sequence of operations that manipulate a separate geometry object.