I've been thinking about #4. I still say that it might be possible to definitively "prove" forgery with the character widths. Consider this hypothetical scenario (I'm pulling numbers out of the air since I don't have a font program to look at the actual character units):
A: 565 units
B: 540 units
...
D: 579 units
I'm talking about *total spacing* here, not just the width of the glyph. A computer can produce a font with any character spacing, but a mechanical typewriter (yes, even a Composer) might not. It depends on how the unit actually moves the paper. There has to be some sort of mechanical item on the font ball that tells the machine how much to move, and I'm willing to bet that it isn't nearly as precise as a computer is. I'm also willing to bet that font balls would have varied by tiny amounts from run to run.
The kicker though is that mechanical machines were most likely *not* accurate to the twip like a PC is. So, for instance, a mechanical tpewriter (even the most sophisticated ones) might have been able to produce a "D" at 575 or 580 units, but not exactly 579 (or maybe they coult make a 579-unit "D" but not also a 565-unit "A"). The point I'm getting at here is that there is *no way* that all of the character spacings would be the same as in MSWord. Some would necessarily vary, partly because fonts from different manufacturers (or even the same one at different times) vary in spacing even if the actual letterforms don't. I also seriously doubt the machining of either the font ball or the typewriter itself would be that precise or reliable with use. What's more, it should be possible to prevail on someone with access to such a machine to relate just how it mechanically moves the paper according to the font spacing. Like the first link showed, the documents were ucannilly precise, not only with each other, but also with a word processor made thirty years later. In my mind, the only way this isn't a forgery is if Monotype bought *this particular typewriter and font ball* from the Air Force and used it as the basis for the Times New Roman we all now know and love in our PCs.
Old mechanical proportional font machines like the IBM Composer had maybe 5 different character widths. Nothing like "twips."
Lower case "i", numeral "1" and lower case "l" would be one unit wide.
Uppercase "w" and uppercase "m" would be five units wide.
Everything else in the middle.
I seem to remember that lower case "l" and numeral "1" were the same but maybe not. I haven't seen one of these machines in years but talked to my old boss about it a couple of days ago.
Brings back ancient memories.