I need help, and right away. (Ain't that the truth, tho...)
Seriously, I could use some help from any interested parties. This site has gotten the much-needed help of someone with an actual working IBM Selectric Composer, and that person as sent back a type sample.
As you can see from their site, there are serious differences between the type sample and the actual memo, which they are using to argue that the Composer could not have produced these documents. However, they may have proven the opposite.
There are several big differences between the memo and sample. First, the line spacing is very, very off; the Composer author admittedly did not set the linespacing the same (it is adjustable to the point, on the Composer.) So that's not really an issue either way.
The superscripted 'th' is incorrect; that's not really an issue, because we have already heard that custom typeheads were available with special characters for large clients, and certainly the United States Military would count as a large client, and one that would want the 'th', 'nd', and 'rd' superscript glyphs.
Most significantly, however, the space allocated to spaces is quite off... the Composer sample has space characters that are roughly twice the size of the memo sample.
That would be a deal-breaker, except for one detail. As the Composer manual, page 18, says:
The Space Bar on your Composer can be used just like the one on a typewriter -- to space between words. Three units is the standard space created by a single depression of the Composer's Space Bar. This, however, can be adjusted by turning the inside dial at the lower right of the machine. Turn the inner knob so that the white line points to one of the colored wedges. The numbers on these wedges represent the unit value of the Space Bar in that adjustment. You can, for example, set this dial at fiave and create a five-unit space with each depression of the Space Bar. (For rapid movement of the Element across the page you can set the dial on nine, the largest number of units which the dial permits you to space.)
So it would appear that the Composer can indeed also create the spacing of words in this document. And note that the spacing doesn't (seem to) vary within the memo; it was just set on one setting, and left there, as you normally would when typing. But that setting is much smaller -- about half -- on the memo sample than on the sample sent by this Composer user, who seems to have the setting on his machine set to perhaps 7 or 8.
We have established that both vertical and horizontal spacing can be duplicated by the Composer, so now there is something much more interesting. The font style of the Composer sample is 11pt Press Roman; Press Roman is the version of Times New Roman developed for the IBM Selectric and Executive typewriters. Look at the sample. Can you see any differences between that typeface, and the memo typeface?
Seriously. This is very important.
As we have discussed, the "Times New Roman" font used to compare the memos by the LittleGreenFootballs crowd was, upon examination, very different from the memo typeface. People have pointed out differences in perhaps a dozen different letters and numbers.
What about this type sample? Can anyone see any differences there? The 9 is a little lower on one vs. the other, but that could very well just be an artifact of the individual machine. The 1's and 3's, as well as the other letters, seem absolutely identical upon my own examination.
What is even more interesting, the proportionality between letters -- take a look at the "F.L.S.", for example, which has rather awkward spacing -- also looks identical in both letters.
I know people are tired of this, as am I, but take a look. Can anyone find differences in typeface, aside from the 'th' glyph? Seriously, don't f--- with me here, look honestly and closely.
Have they just inadvertently proven the typewriter used to produce the Killian documents, down to the typeface, point size, and settings?
UPDATE: And note that the IBM Executive, a much cheaper machine, was available with this exact same typeface (though we don't yet know the spacing characteristics.)