Drawing GUI Elements

Hmm. By delving in to the concept of a 3D GUI, you're looking at some of the basic principles of "3D graphics" (at least what I think the basic principles are).

Thing is, on a flat screen everything's just a simulation of "3D" in any case; the question is: do you go with a complicated interface to your display that lets you input geometrical coordinates and the like in 3 dimensions, just to have it all turned into 2D in the end, or do you simulate it yourself with bitmaps and the like?

Now (and this is all merely my opinion), something like a GUI is not going to need to be especially realistic; therefore it would probably give better results if the 3-dimensional-ness illusion was programmed straight.

Here's where I'm not especially certain about where you're going with it: how much of a "3D look" do you want this to have? I guess you might need to describe what you want it to look like a bit more. I mean, is this all going to be modeled onto different sides of a cube, or a sphere, or what? (Or am I missing the point entirely?) Since people can only see one side of something at a time, it usually makes a lot more sense to just model that side, and in 2D (again, since it's all going to end up on a flat display anyway - unless you're working with top secret holographic technology or something).

So I have a question in return: what will your polygon containers be like, and how will users navigate this 3D GUI?

Twilight Dragon

(slap) How did that happen? Could someone get rid of that extra post, please?

[edited by - TDragon on December 9, 2002 4:28:43 PM]

Okay this is the premise behind the user interface.

Take your typical ''3D world''.

The user will be able to at any point in the world call up interface panes pertaining to an element on screen. e.g. if you are looking at a sphere you can click on the sphere and bring up a window describing its properties. Now if you move back the pane will remain where it was originally called so it will shrink. If you rotate your view side to side you will see the pane at an angle.

A simple analogy is that a window is like a post it note. If I stick a post it note on my monitor when I move my head the post it note stays and place and does not move with my view.

My basic thinking was render the window as 2D then draw it on to a polygon.

Oh using the interface in 3d is nice and easy...

The mouse click is simply converted in to a line and all the components that collide with this object are considered. The one closest to the screen is the one activated.

My idea for moving the interface is quite simple too, you simply move to the position you want the window and drag the window title bar (or select the move option). This sticks it to the current 2D plane aligned with the screen and it acts like a normal window being moved.

Ok, I get it now. That sounds like a reall great idea.

My advice would be, yes, do it as a texture. Create the texture on the fly using functions to create buttons, menus, etc. Create a rectangular 2d surface and texture the panel to it. Your idea for moving sounds like it''ll work fine; just move the surface according to the mouse movement and the texture-map info panel will come with it. You will have to re-draw and re-map the texture every time something changes, but it hopefully won''t be to much of a drag-down.

Twilight Dragon