A Borg ship normally wouldn't be a holiday destination, but yeah.
Mandelbulber does
very nice renders, but so far
none of the software I've tried has an ideal navigation interface. Ideal, I think, would be a) clicking took the camera viewpoint halfway to the nearest part of the fractal where you clicked -- the z-buffer theoretically allows this, or a single raytrace could be done at the clicked pixel -- and rotated (first horizontally, then vertically) to center that point in the view, and b) the camera could be rotated in place, without moving it. As near as I can tell, Mandelbulber has a target point and puts the camera "zoom" units from it in the direction determined by the rotation, so changing the angles actually moves the camera, swinging it about on its invisible tether, and click-zooming sometimes has unpredictable results since it uses the DE values instead of the actual, measured distance to the specific fractal point clicked on. All the others' navigation is, if anything, worse. So, a constructive suggestion for 0.44: the xyz coordinates become the camera position and the angles set its facing direction. A third, roll rotation that keeps the point in the center of the view and is applied last might be nice to add, too. The zoom would continue to affect the scale of fog and DOF effects and near-plane clipping; the default after a click-move could simply be the distance to the fractal point targeted. Other nice-to-haves would include a gradient editor, ability to animate parameter morphs and not just camera movement, ability to add a light source near but offset from the camera (distance left-right and distance up-down from camera, scaled by zoom), and ability to generate a stereo image pair (separation parameter, multiplied by zoom, determines two viewpoints left and right of the current camera pos). Movement buttons to strafe the camera or march it forward or back in steps of zoom/10, perhaps, too. This would yield much more "first person shooter" type control over the camera.

Right now I'd say the two big weaknesses though are the current navigation model and the trial-and-error nature of generating color schemes.
Despite these, I'm producing some remarkable results it seems.
