Help me understand zoom vs Z elevation

Me and a couple of friends are using Construct for my first major game project, and although v0.9x isn't perfect it really is helping us create this game quickly and efficiently. Here's an example CAP of the first level, with preliminary graphics: http://dl.dropbox.com/u/3877194/ILostmy ... review.cap

However there are some areas in which better documentation would be great. I've had trouble finding info about layer zoom and z elevation.

In the example you can see that I'm going for this pseudo-3D parallaxy thing using zoomed layers. Afaik, the only difference between zooming layers and changing Z elevation is that objects of a zoomed layer never go behind or in front objects on other layers, while objects changing Z elevation can do just this (with the 3D layering option checked). Is this correct, and am I using zooming the "correct" way? I took this from the Spaceship Demo, and it would be great to know before making a lot of levels.

Secondly, is anyone aware of the bug where shadow caster shadows are offset on sprites or tiles? See attached picture. Top box is a tile, bottom is a sprite. However, I later had the same problems on sprites. Tried changing hotspot placement but that only fixed it sometimes. Very strange. In the end I had to create a separate shadow caster sprite object, the one marked with a black cross in the CAP, for this to work reliably.

<img src="http://img560.imageshack.us/img560/6118/shadowcasterproblem.png">

Thanks in advance!

Z elevation relates to the camera position, where zoom goes to where your x/y position is on the layer.

Basically Z height is a 3d representation saying that if this object is on this side of the camera offset the object a certain amount, and then offset some other amount if its on another side.

And then if its right in front of the camera don't offset it.

All of that based on a percentage .

Edit:

No idea whats up with the shadow caster.

newt: Thanks for the info. However when not zooming to a x/y-point (akin to what I'm doing with MagiCam in the demo) but rather zooming a whole layer it seems to be pretty much exactly equivalent to changing z elevation for all objects on that layer, with the parallax effects you describe.

I guess this means that unless want objects from one layer moving behind/in front of objects on another layer I may as well just use different zoom on different layers. I just hope it doesn't clash with some special plugin use-case.