in Poser 7:

I've never used colliosion detection before, so I test it out. Sure, it stops a ball at the ground. 
But, it seems to be nullified by the "tweens" between poses.  I start with Daz Pup "Lay Flat" pose on frame 1, then on frame 30  apply the "Lay Out" pose.  In going from one pose to the other, the front legs are NOT prevented from penetrating the ground, even though the collision detection has been properly applied to Pup (hip and children) and Ground, and collision turned on.    
Any Ideas?

I believe the collision detection is only active while you are posing.  In other words,
it stops collisions when you're setting up frame 1, and when you're setting up
frame 30, but it's not active "during the animation".  If a spline pushes the parts
through each other between those two frames, the detection won't detect.  :)

I detect a lot of wit i what you are saying, and all of it, as usual, is quite helpful thank you! That explains it. 
Yeah, okay that's good enuf, I can work with that, I suppose I don't need "tweening" detection between poses. I guess.

I quit using the feature after I understood its limitations. 
When posing, I can detect collisions well enough by eyeball.....
an automatic stopper would be very useful during the animation,
but that's not when it works.

Dr. Geep has a whole tutorial on collision detection and definitely it's only practical use is in posing only.  If memory serves me correctly, CD is only limited to several selected areas, not all.  Plus it bogs down render process considerably.

Thank you both for the very helpful responses

It is one of the last great frontiers. I hope someone can eventually fix/solve it (CD).

It doesn't seem like it would be a far reach from the current capabilties. I mean, you've got frame by frame collision detection when you do cloth (or hair) dynamics... see what  i'm saying?  Or maybe I'm confused, but it seems to me that is what is happening... if you can do this for something "soft", why can't you do it for something "hard"? 

Yup, that's the appropriate question.  There are quite a few ways to
do collision detection; some are quite fast but not precise.  Those
could be incorporated into the animation code.

The cloth and hair rooms were built by other companies and connected
into Poser.   The animation code is an internal part of Poser, and (so far anyway)
none of Poser's owners have seen fit to make any important changes to
internal code.  Except for the Layers bit, which doesn't work right, there
haven't been any improvements since P4.

Quote - in Poser 7:

I've never used colliosion detection before, so I test it out. Sure, it stops a ball at the ground. 
But, it seems to be nullified by the "tweens" between poses.  I start with Daz Pup "Lay Flat" pose on frame 1, then on frame 30  apply the "Lay Out" pose.  In going from one pose to the other, the front legs are NOT prevented from penetrating the ground, even though the collision detection has been properly applied to Pup (hip and children) and Ground, and collision turned on.    
Any Ideas?

Somewhere around here is a python script that constrains body parts from poking through the ground during animation.  ie: Figure's feet/etc will stop at groundplane.  I don't do any animation so I've never had a use for it.  Usually, I have scripts like that stored just in case I might need them.. but, I couldn't find it.

The point is:: There is a python script somewhere that someone did which helps with that.  Check here and Sharecg searching for python scripts/animation etc.

My "Walk Animation Script V2" here in freestuff has a very limited version of in-animation collision detection. It's actually based on ockham's WalkThisWay scripts.

I think this shiould be doable using PyODE and PoserPython. I don't think this kind of functionality is included in PoserPhysics, though I could be mistaken.

Hmm, a per-frame callback, simulating the movement from the previous frame to the current frame, using ODE to check for collisions (or maybe NumPy)... Quite a programming challenge, but not undoable. The results wouldn't be precise, but they'd be good enough for animation.

By the way, collision detection has absolutely nothing to do with rendering, and turning it on or off will have no influence at all on rendering speed.
It WILL have a serious impact on the responsiveness of the user interface while posing - in fact, reducing it to unacceptable levels. That's why I don't use collision detection as implemented in Poser.

cloth collision needs a rewrite too.  maybe some third party could write an alternative for it?  i won't hold my breath though...

It WILL have a serious impact on the responsiveness of the user interface while posing - in fact, reducing it to unacceptable levels.

yep, confirmed that one.

(slightly expanding the topic):
You know, there is a lot to love about Poser, I've been getting won over these last few months doing animation. I now actually LIKE and appreciate how keyframing / spline wrangling  can be done there.  I just have 3 biggies left in my wishlist for improvements: - export 64 bit RBG (16 x 4 channels)

• animation layers that actually work
• collision detection in animation

There's collision detection and then there is hard/soft body collision physics.

The first just keeps parts from going thru the mesh of other parts. It has it's place.

But true physics, where after collision the objects colliding bounce back (react) in a realistic way, that's another. On top of that, for soft body the mesh of one or both objects must deform on collision.   The artisit must be able to control the settings of rebound, amount of deforation, etc.,

Lightwave:

http://www.youtube.com/watch?v=SA221laaaG0
http://www.youtube.com/watch?v=wKH9T1tpt4s&feature=related

::::: Opera :::::