QUESTION
What's an arcsecond? 

ANSWER from Sylvia Baggett on June 21, 1996:
An arcsecond is 1/60th of an arcminute, which is 1/60th of a
degree (so the arcsecond is 1/3600th's of a degree). To give you a
feeling for the size of these units, the following examples may help:
the apparent distance between the 2 pointer stars in the Big Dipper
constellation (the ones that point to the North Star) is about 5
degrees (or 18,000 arcseconds) while the diameter of our Moon as seen
from Earth is only about 0.5 degrees (1800 arcseconds).

If you prefer a more down-to-earth example, have a friend hold up a
coin the size of a half-dollar (about 2.5 cm or 1 inch across) and
stand about 1.5 meters (about 5 feet) away from you. The apparent size
of the coin from that distance is about 1 degree. To have the coin
appear only 1 arcminute across, your friend has to stand with the coin
about 86 meters (286.5 feet) away from you (almost the size of an
American foot ball field!).  And to have the coin appear only 1
arcsecond across, your friend with the coin has to be about 5160 meters
away (almost 3.3 miles!!). As you can see, it's difficult for the human
eye (without any binoculars, telescopes, etc) to see things 1 arcminute
across and impossible to see things 1 arcsecond across.

As for the shades of grey that are supported, that too is camera
specific. In WFPC2, for both WF and PC, we have what is called a 12-bit
A-to-D converter, that is, a piece of electronics that takes the analog
signal detected from each pixel and converts it to DN, or numbers. The
12 bits limit the DN you can have to 2 raised to the 12th power=4096;
numbers higher than 4096 are not possible with the WFPC2.  So, you
could think of the WFPC2 as having 4096 possible shades of grey.

As for the FOC, it is a "photon-counting" device instead of an analog
device like WFPC2.  In other words, the FOC actually counts the arrival
of each photon as it is detected, so there is no A-to-D converter
involved.  As a result, the number of shades of grey which the FOC can
detect is limited by how much memory each pixel is given, and can be
either 8-bits per pixel or 16-bits per pixel depending on the details
of the exposure.  An 8-bits per pixel image can distinguish 256 shades
of grey (2 raised to the 8th power), while a 16-bits per pixel one can
see 65536 shades of grey.