The potential of digital cameras for astrophotography under light polluted skies may be easily overlooked.
My camera is a Canon S400 (no longer marketed) and it has various ISO approximated speed settings as well as the ability to take long exposures. I took it to a dark sky site 60 miles (about 100 km) from Grand Junction, Colorado to photograph constellations. Then I photographed the same stars from the city.
The dark sky site has a naked eye limiting magnitude of about 6.2 while in the city it is 4.6. The limit reached by my camera (at 'ISO' 200, 15 second exposure, f2.8) was mag. 8 under both dark skies and in the light polluted city! Perhaps the background sky glow in the city is not recorded as well as the eye sees it (at least at the above settings).
The following photos were not edited except for size adjustments so as to fit this webpage better. They are 150 pixels/inch. The limiting magnitude is the same for both photos but the contrast is quite different. Your monitor settings will influence how many stars are visible (decreasing monitor brightness limits stars on the dark site photo more than on the city site photo). Also, LCD screens, but not CRT's, seem to favor more stars in the city photo.
Scorpius: dark sky (L) and light polluted sky (R). Photo limiting mag. 6+. Scorpius is near the southern horizon at my latitude.
Lyra: under dark sky (L) and under city sky (R). Photo limiting mag. 8. Lyra was almost at zenith.
A digital camera might perform better in light polluted skies than a film camera. I could not find data on the spectral response of my digital but I know that it picks up near infrared (I can see the light of an infrared remote control on the camera's LCD screen). If at the same time the near UV response is decreased (and I do not know if it is or isn't) then the digital camera won't pick up on the scattered light pollution in the sky from shorter wavelengths (mercury vapor lighting) as much as a film camera might. (The sky scatters more light at the blue end of the spectrum than at the red end).
Comparing the response of film and digital is not exactly fair. The light sensitive medium of each are quite different as are the types of settings available. Of course filters (built-in, as in the case of digital cameras, or external add-ons) greatly effect limiting photographic magnitude. It would be interesting to find out the limits of each at their optimum settings. At the present I have no time to pursue the digital/film, light pollution/limiting magnitude issues but there are a number of astrophotography websites that can be researched. I don't think that any common non-cooled digital camera can beat good old film for really great constellation photos. Color slide film (ISO 1000) and Kodak's T-Max b&w negative film (processed as slides and pushed to ISO 3200) give wonderful results at only 30 seconds exposure.
Update October 2004
I took photos from a dark site with the same camera as mentioned above. Settings were an approximated ISO of 200 and 400. The constellation is Grus (diagonal line of stars in lower right corner). Bright star in upper left is Fomalhaut. You can decide which is better - the 200 or the 400 settings. By the way, both were 15 second exposures at f2.8 (my wide angle setting). I did experiment with a three-power telephoto setting but that made the stars too dim and almost none could be seen on either ISO setting.
In each right photo I used Paint Shop Pro 7 to adjust the histogram by expanding the mid-tones. Take note of how much brighter the stars are with this one simple adjustment!
Grus - (L) '200' ISO; 15 seconds. (R) Same photo but histogram adjusted.
Grus - (L) '400' ISO; 15 seconds. (R) Same photo but histogram adjusted.