For some reason my brain just won't comprehend the concept of object space vs world space vs view space vs screen space- please help!! Actually, I understand what each of these terms mean and how they relate to each other, just not practically how the concepts are implemented.

All of these questions relate to the design of a FPS (probably an unimportant fact).

First Question -- I want to design several rooms joined together. Should all the polygons that define these rooms be in world space (which means to me that all the polygons are relative to the same origin), or should the polygons be in object space (which means to me that each polygon would have its own origin)?

Second Question -- Now that I have several rooms, I would like to put a camera in one of them (a.k.a. go from world space to view space). How do you figure out what the initial transform will be so the camera is in the middle of one of the multiple rooms you've created? Or is it better to design the rooms initially (assuming the answer to question 1 is that all the rooms are designed at once in world space) with the origin being where you initially want to put the camera?

Third Question -- Now I want to place an object (i.e. a crate, etc) in the corner of one of the rooms. I am fairly sure that the object should be designed in object space (which means to me that it has its own unique origin) so that the object can be moved, rotated, etc independently of the world. Once again, how do you figure out the initial transform so that the object end up in the corner of one of the rooms you've created?

I am assuming that the rooms will be created using some type of level editing program.

Please help before my brain explodes.

Thanx

I may be a little off base, but here it goes:

First off, object space does NOT mean a unique origin for each polygon. That would be insane. Object space is exactly that: space that an entire object is defined in. That way, one transform will let you place the object into "world space."

I don't think anything in your game is defined in terms of "world space", just large object spaces. For example, your two rooms, assuming they are static, are basically the same object, and should be treated as one. In fact, an entire static map should be treated as one object, and placed into "world space" like every other object.

Using the above as a model, it becomes clear that the crate is defined in its own object space, and is placed into the "world space" using yet another transform.

As for the camera, generally the camera is placed in the world, not the world is designed around a camera.

Checking out the articles on the introduction to 3D in the Direct3D documentation would help a lot. Many complicated concepts are descibed in better ways than I can (read: with graphics )

- Splat