I wonder has anyone done any testing of the same picture in these different methods?

Not for the purpose of generating a file for printing, or capturing all the information in the negative, but for quick digitizing to share on a website.

It seems to me like flatbed scanning a print, even using weights like books to get the print as flat as possible on the glass, (Mostly talking 4x5 fiber contact prints here) would likely not result in as high a quality as using a piece of museum glass and rigging up decent lighting and a good macro lens, both in terms of flatness, and in terms of picking up less surface texture of the paper.

I guess an important distinction to make is that I don't want to do all the burning and dodging over again in the computer...this is for negatives I never scanned before; if I was going to reprint with digital output obviously I'd make a proper scanner scan or camera scan from the negative...but just thinking more in terms of speed and efficiency and quality just for the screen.