[stan_rogers]

The authors of the paper seem to believe that the object would have been to slow down the typist (had the jamming hypothesis been correct). The opposite would have been true. Increase the difference of the angle of approach of successive hammers, and you decrease the chance of jamming and increase the speed. I'll admit that the Z-SE ("dot-dot-dot-space-dot" in early American Morse, with the length of the space/mark mkaing the difference) problem makes some superficial sense, but then there's a corresponding R-EI ("dot-space-dot-dot") and others that aren't reflected on the QUERTY keyboard. (The original Morse code needed far too much space-length discipline, which is why it was replaced fairly quickly, even in the NIH-heavy USA.)