Mah'scope -- ha! that's great..
Well, about the two pics below:
The cmos chip that grabbed the image was 1MP 960x1280, and then digital-zoom (what a marketing gimick!) reduced the resolution to 120x160, a factor of 8! Look! it correctly assigned grey pixels according to where the craters are..
As you can plainly see, interpolation with digital zoom does not make up for optical zoom. But little tricks like using natural noise and camera shake to reposition the image over other ccd elements you can "fill in" the pixel that was supposed to be between the other two pixels (or ccd elements). A good algorithm will generate sub-pixels based on rate and intensity change between frames or camera movement. Kinda like a microscopic parallax analyzer, but with a different goal in mind. This is different than just upscaling your image in photoshop.
By 2010 or sooner we are going to see 20Mpixel imagers, with 12-15Mpixel chips being the norm in professional cameras right now. Pocket cameras nowadays are up to 9Mpixel, with 6-7Mpixel being the sweetspot right now. I'm talking consumer-class products here now because the largest CCD array today is 84MP !! OMFG!! ..and is intended for research applications.
http://luminous-landscape.com/forum/index.php?showtopic=3598 This has got to be able to resolve far more detail than film! I think the elements are smaller than the halide crystals! By 2x
I bet the guy who invented photography is flipp'n in his grave right about now
One of these pix is from my old trashy fone and the other from the voyager spacecraft.