The stress thing I get, though the level of forces induced by anxious fingers vs. friction and springs is probably doable for any well made shim. Losing the parts is valid. Perhaps is could be like getting individual cassette cogs -- buy a Campy 2004 9sp ring for $13.99 to go with your changable shifter.

The point of variable pull for different cog pairs -- 3.185mm from 11T to 12T, 3.190,mm from 12T to 14T ... -- seems to indicate an elliptical drum/shim profile or even a multi-elliptical one. That would indeed be a ***** since angular registration of the shims on the drum would need to be within a couple of degrees. That would imply that instead of end-user replaceable shims, the part would have end-user replaceable barrels.

For shxts-and-giggles, I softened some copper electrical conductor and squished it into a V shape using a vise and file. Wrapping that in between the shift cable and the trough in my Suntour barcons basically worked. Indexing went from perefect to non-existent, skipping about 1.333 gears per click. So, though that was an utterly useless proof-of-concept, it works for the simple stuff.

Perhaps I need to get the LBS to let me look inside a couple of different brifters. I'm never going to afford "experimental" sample brifters.

Thanks for the info. Pointed to areas to learn about.