cwstjdenobs said:
Why not just have the whole image in focus all the time?
Because depth of field is a neat and cool effect, and a picture where everything is in focus would look flat
Plenoptic cameras can give you uniform sharpness across the entire picture (within certain boundaries; you can't get sharpness across an infinite depth range), but that's not the point. They record the light field, and you can easily emulate different focal lengths, different lenses (wide to narrow, orthographic to fish-eye), or even move the viewer forward or backward, left and right, and extract (slightly) 3D images, from the data contained in one image. But adjustable depth of field is the most striking of those features (and the one with the greatest range allowed by the technology), most effective, and no wonder Lytro chose to focus (heh) on that most of all.
Not sure what other features they'll decide to implement. Of course they'll enable having the entire picture in focus, but playing with the focus and focusing on individual layers is pretty interesting and will probably be the major draw of the camera.
Here (see bottom of page for PDF file and video link) is an explanation of the principle, some examples, and a video showing the capabilities of the tech. It's pretty interesting.
One major drawback is that the end photos will probably be in relatively low resolution. Because of the way the technology works (an array of microlenses over the optical sensor, each covering a grid of pixels; the end photo is composed by extracting one pixel from each lens), the end photo has a significantly smaller resolution than the optical chip itself; let's say each lens covers an 8x8 grid, that means a 10 Megapixel optical chip would only give a 456x342 image.
EDIT:
[youtube]http://www.youtube.com/watch?v=9H7yx31yslM[/youtube]