Re: [Rendering order] z-depth and 3D effects

Olaf,

At 18/11/2009 09:54 PM, Dr. Olaf Hoffmann wrote:
>Hello,
>
>I think, the main problem with a 'real' z-axis is, that
>for 3D-transformed objects there can be a different
>z for any fragment of the object.

That's true for 3D. But we're only talking about 2.5D, so 
considerable compromises must be made. IMHO, the first major 
compromise is to treat a graphical element as existing at one and 
only one point in the z dimension, even when transformed.

>Therefore if there is more than one object, the
>objects can intersect in a complex way.

Treating objects as single z points (flat) prevents them intersecting 
in complex ways.

> From a simple 2D-viewer one cannot expect
>a proper computation of such intersections without
>the help of the authors. From a 3D-viewer one can
>expect this.
>
>If SVG2.0-viewers are not expected to be such
>advanced 3D-viewers, objects can have only one
>z-value to rearrange the rendering order without
>solving the problem of intersections.
>If this z-information is explicitly given or at least
>implicitly known to the viewer for each object,
>such a 3D-effects model can be kept simple
>(there can be of course different approaches more
>or less useful/convenient for authors).
>
>The first step to clarify the intended 3D approach for SVG 2.0
>therefore would be to say, whether a shape/object/path
>can have more than one z-value or not.
>After this decision or clarification one can go into details
>how to get any effect at all.

I agree.


>Olaf

Steve.