HP 740: It saves room on the front panel. As you can see, there is none to spare - they even put the input/output binding posts on cables.
I wish they'd allowed you to turn the knobs from 9 to 0 instead of forcing you to crank all the way CCW. Dialing it up and down to follow a varying input is clunky, probably a reason differential voltmeters were discarded when DVMs came on the scene. But they can do things you can't do with a DVM and that's why I keep them.
Monsanto 2000: I rediscovered the Monsanto article, actually only the abstract because the American Chemical Society wants $10 for the full article.
https://doi.org/10.1021/cen-v044n004.p048Restoring it would not have been possible without the service manual which I was lucky to find.
Fluke 8100: I don't know why they restricted it to 12000 counts, they could have done 16000 with virtually no changes.
UPDATE: A bit more than no changes, because the primary DAC is weighted 4-4-2-1 instead of 8-4-2-1. The predecessor models counted to 11xxx... so 12 may have seemed like overkill.