This is a very interesting question, so I decided to do small test. I just printed off some very small black and white squares arranged like a chessboard. There are 16 of these squares per centimetre on a small section of the paper. I supposed that if resolution is thought of as "the ability to distinguish between two points" then all I will have to do is to walk away from the paper until I can't tell that the squares are not just one solid colour. Admittedly, this is a pretty crude test. Added to that I have been staring at a computer screen for over an hour and it's quite dark, and I already
know there are squares on the paper so that's what I am expecting.
Right, I've just done the test once and the squares started to form a sort of speckled grey at around 2.8m from the paper. In terms of angle, this is a resolution of about
0.013�, or � of an
arc minute (one arc minute is equal to one sixtieth of a degree). This figure, while not accurately measured, ties in quite well with the figures given on the site below.
The site also quotes the
fovea (the high resolution part of the eye, the part that sees in the best quality) to have a total of 24 million pixels. If I understand this correctly, this means that in any one direction it can distinguish between 2800 points, assuming the fovea to be roughly circular. As for its resolution, I have learned from a none-too-reliable source that the fovea is 0.2mm in diameter. This gives it an approximate resolution of 13 500 pixels per centimetre, or 35 000 pixels per inch. (Please don't take any figures I tell you to be accurate.)
http://www.opticalphysics.com/Vision.htm