Maintaining the internal radius should be standard practice on any crank re-grind. I don't know why it is emphasised in my AJS/Matchy manual. Its not as though the average Joe will do this kind of work at home.
Ha! If you'd had a crank break on an AMC twin you'd wish the radii data was in
BOLD TYPE 12 feet high! Radii matter on everything, dead right, and decent engineering firms handle them correctly. But, and here's the sad bit, quite a few don't, or didn't.
When I broke a crank (on a 500 amc twin, not a roadburning tuned beast) it was a mere 40 miles after a grind and a lot of expense. Classic fracture on one end of a crankpin, causing a fair amount of fresh expense. Now, those three bearing engines are a bit rigid down below, and critical oil passages pass quite close to the surface in the area of the centre main bearing, which maybe explains why the AMC literature is spattered with warnings, but the same principle applies to anything. When Austin introduced the better-supported three bearing crank on the late '7' in the 1930s . . . they broke alarmingly frequently compared to the bit of bent wire that passed for the 2 bearing item! Were they just too rigid, or was there a design flaw - I haven't a clue - but many period engines with twice the power of an Austin 7 ran perfectly happily on single-plane 2 bearing cranks.
Thing with the radii is, if damaged or absent they can't be put back, the structural integrity has gone for ever and all you've got is a doorstop. Good engineering firms will refuse to work on anything with damaged radii, as it's the first thing they'll look at. But there be others, reputable and no doubt good at most of what they do, who don't always appreciate how marginal some of these older designs are, and how susceptible to breakage.
I have been the unproud possesssor of a few doorstops after buying bits unseen, or assembled and uncheckable on the spot . . . . . .