After looking closer at the patent skectches for the Model M switch, it looks like we could easlity shorten the travel to the actuation point, but maintain the overall travel, without much work at all.
All we'd need to do is add 1 or 2 (maybe 3) washers UNDER the spring to precompress it a little. They'd have to be the same ID as the post on the hammer. The force/moment geometry is unchanged so it'll behave the same way on the downstroke, just trigger earlier.
BUT....
Due to the designed-in changing upper end connection on the spring (fixed -> pinned -> fixed) we'll be limited in how much shimming we can get away with before the spring won't "unbuckle". That upper end connection is key to why the "make" point is so significantly different than the "break" point.
Rip, do you still have that bolt modded Boscom function row handy? That'd be perfect for this experiment. My only M isn't bolt modded yet.