My new Sony DSC-W300 seemed kind of flakey, and after much fussing around, I came up with a very simple and fast test which confirmed it was indeed a dud. Here's how:

With camera in widest-angle setting and wide open (if possible) photograph trees at infinity, taking care to fill the corners of the frame as much as possible. If trees aren't available, find something else with plenty of detail and contrast, maybe telephone poles/wires.

Now shoot the very same scene with the camera flipped upside-down.

If time allows, repeat at other focal lengths as well, but it's the wide settings that are likely to be the most revealing.

Back at the computer, turn photo #2 right side up and compare the two at 100%. Is image quality roughly the same in all corners, or are some much better than others in terms of focus, contrast, blue fringing etc?

In the case of my Sony, the left-hand side looked a lot worse than the right, so I exchanged it for another unit which is much closer--not perfect, but near enough that I'm not going to see any real differences at sane magnifications--and with this camera, I can more or less reproduce the image quality that Michael Reichmann praised so highly.

In retrospect, I probably should have tested my Fuji F50fd in this same manner. Instead, I fussed with the thing for weeks and wondered why the results were so inconsistent, and by that time, it was too late to return.