why would anyone still use film?

I've completed Phase One of my master plan to conquer digital photography:

1) Make sure the equipment works.
Pre-req: Must stop buying new equipment at some point.
Pre-req 2: Must stop dropping equipment at some point.

2) Learn how to use the equipment.
Progress: Another month or so to go on this one.

3) Learn how to take good photos.
Progress: The hardest one. I think I'm making progress though.

What's great is this stuff is so much easier to learn than all that film crap people put up with for 150 years or so. With the Digital Rebel, there's basically only a few things you have to get right (I love lists):

1) Subject
2) Composition
3) Lighting
4) Depth of Field

Everything else is either fairly reliably automated in the camera, easily corrected after looking at the preview, and/or fixable in Photoshop (within reason). This includes focus, exposure, color balance, saturation, noise reduction, and developing. And with enough fiddling in Photoshop you can even fix the Big Four.

(Photography is so addictive. Do yourself a favor: unless you're a graphic artist, never install Photoshop on the same machine you're trying to get work done on.)

Furthermore, I can imagine a day when Depth of Field will be a non-issue. If someone were to develop a fast rangefinder that could take as many samples as there are sensels in the imaging sensor, then the rangemap could be used to choose focus in post-processing. The issue then becomes compensating for noise caused by either diffraction from the tiny aperture or very small sensor sites required to get good depth of field.

For that matter, with a "z buffer" as described, lighting could also be manipulated in a more convincing way. Of course, this means image manipulation must also now occur in three-dimensions. But computer game artists have been doing this for decades, so it's certainly possible.

Finally, subject and composition, while not "fixable" with this technology, will at least become more flexible in post-processing. It will be trivial, for instance, to slip your Uncle Ned into the group photo by merely assigning him a distance intermediate to the people in front of and behind him.

One approach to this technology would be to adapt DLP chips. In one orientation, the DLP mirrors bounce light to the imaging sensor. In the other, they reflect one or more lasers out the lens (and back in). For instance, a single laser can be oriented to point along one of the edges of the DLP array. A row of mirrors (one for each column in the imaging sensor) turns the beam 90 degrees down one of the imaging sensor's columns. The sensels turn one by one, and the range is read off. This must be done millions of times before each shot. For still subjects, "focus" speed is not that important, but this will be a challenge for action shots.

Another drawback is that the mirrors will reduce image quality. But there are already SLRs that use a fixed half-mirror to avoid mirror slap, so it's not unfeasible. On the plus side, you can remove that bigass SLR mirror that makes your camera shake every time you take a picture.

Coming soon: thoughts on alternatives to lenses