Depth of field discussions: The factors:

1. Size of the circle of confusion.
2. Focal length of lens.
3. Aperture setting of lens.
4. Distance focused upon.
5. Degree of enlargement to render displayed image.
6. Distance observer to displayed image.

The roots of this discussion: A person with 20/20 vision viewing a disk from a distance 3000 times its diameter sees a dimensionless point. The shape of the object is unidentifiable. Now 1/3000 the distance works out to 1 meter diameter viewed from 3000 meters. However, for photographic purposes, due to optical flare, film turbidity, grain structure, and noise, we use a far lower standard. For photographic work the angle sustained is 3.4 minutes of arc. We are talking 1/1000 of the viewing distance as seen from 10 inches This is a circle size of 1/100 of an inch (0.25mm from 250mm).

If the image is viewed from 20 inches, this works out to about 2/100 of an inch viewed from 20 inches (0.5mm viewed from 500mm). This is the stuff of depth-of-field tables.

Now todays miniature cameras yield an image that is mostly useless unless enlarged. To get an 8x10 image from a full frame (24mm by 36mm) the degree of enlargement is about 8X. Thus the size of the circle at the focal plane of the camera must be 1/800 of an inch in diameter or smaller (0.032mm).

Based on all this gobbledygook, it is common practice for the industry to use 1/1000 of the focal length to construct depth-of-field charts and tables. The Leica standard is 1/1500 and the Kodak standard for precision work is 1/1500 of the focal length.

I think camera makers could adopt 1/1000 of the focal length for the size of the circle of confusion and build auto DOF routines. They could make the circle size selectable by the user.

Just my opinion Alan Marcus

Depth of field discussions: 
The factors:

1. Size of the circle of confusion.
2. Focal length of lens.
3. Aperture setting of lens.
4. Distance focused upon.
5. Degree of enlargement to render displayed image.
6. Distance observer to displayed image.

The roots of this discussion: A person with 20/20 vision viewing a disk from a distance 3000 times its diameter sees a dimensionless point. The shape of the object is unidentifiable. Now 1/3000 the distance works out to 1 meter diameter viewed from 3000 meters. However, for photographic purposes, due to optical flare, film turbidity, grain structure, and noise, we use a far lower standard. For photographic work the angle sustained is 3.4 minutes of arc. We are talking 1/1000 of the viewing distance as seen from 10 inches This is a circle size of 1/100 of an inch (0.25mm from 250mm).

If the image is viewed from 20 inches, this works out to about 2/100 of an inch viewed from 20 inches (0.5mm viewed from 500mm). This is the stuff of depth-of-field tables.

Now todays miniature cameras yield an image that is mostly useless unless enlarged. To get an 8x10 image from a full frame (24mm by 36mm) the degree of enlargement is about 8X. Thus the size of the circle at the focal plane of the camera must be 1/800 of an inch in diameter or smaller (0.032mm).

Based on all this gobbledygook, it is common practice for the industry to use 1/1000 of the focal length to construct depth-of-field charts and tables. The Leica standard is 1/1500 and the Kodak standard for precision work is 1/1500 of the focal length.

I think camera makers could adopt 1/1000 of the focal length for the size of the circle of confusion and build auto DOF routines. They could make the circle size selectable by the user.