Don't want to flog a dead horse Bike Beesa, but I don't think it's quite fair to lay the problem at the door of morons using duff bolts. There actually is an inherent flaw.
Even fastidious owners can't tell in advance when the blessed thing is not going to extract the gear- whereupon the bolt can't be got out for examination or replacement. It only takes one very tight fitting taper to wreck the start of the lh thread, and then you're where Simon is. (Even when you have the unit off the machine and on the bench it isn't often that simple to get a damaged bolt out.)
I've got a few damaged ones lying around, which I keep as models for turning up replacements - so it's a problem I've seen a lot more of than I would ever have wished.
The principal of the auto-extract is perfectly fine, elegant and well- recognised, I completely agree. Excellent idea. It's the detail execution of the particular item that isn't so fine or elegant. My own drives for attaching mags of varying diameters, tapers and threads to bench test equipment all self-extract - but they use proper internal shoulders. They are used a dozen times a day sometimes and they don't fail, even on the big tapers found on a lot of 4 and 6 cyl mags. They all just need the one 3/8 BSF spanner from any ordinary toolkit, nothing fancy, same as an atd unit.
The number of times Simon's problem arises is too high, I think many would agree. My heart is always in my mouth when taking the things off machines, with that 'will it or won't it' worry. When it will, all well and good and then one can check the thing for condition, replace or whatever; when it won't, it can be a lot of time wasted. My point is not that the concept is bad, it isn't. But nothing will ever persuade me that is a respectable engineering approach to use the first turn of any thread as the load-bearer to get a tightly mating taper apart!
I can see where you and Simon are coming from, but if a part has worked fine for 60 years then there is not an inherit design flaw in it.
That does not mean that it could not have been done better, but it does mean it was done good enough for the purpose.
Design flaws are the problems you read about in the Agony Aunt columns of PERIOD PUBLICATIONS and the self extracting bolt never rate a mention in the magazines of the 50's 60's 70's & 80's.
It is a bit rude to criticise a design from the 50's 60 years latter when it was not a problem in the day.
Exactly the same applies to stiction between the slides & carb bodies and wet sumping.
None of these were a problem when the bikes were being used within their design parrameters.
I ran 2 A 10's in very wild youth, one a fire breathing monster 11.5:1 running on a nitro benzene/ petrol mix and the other, a bog standard rat bike A 10 that I hoped would still be where I left it after lectures.
The only problems I had with the magnetos was ripping all the teeth of the drive gear on the Friday night drags along Friendship Rd Port Botany.
On both the bikes, pulling the magneto pinion was a weekly job so if I pulled one off I pulled a thousand off and never ever was there a problem.
Fast forward 30 years and every second idiot has got an impact driver 4 times the size of their IQ and they think it is perfectly fine to rattle a magneto onto a taper or smash it on with a 4lb hammer when all that is needed is a gentle squeeze, and this is where the rot sets in with tapers jambed on way too tight so naturally when some one tries to extract it it is beyond the shear strength of the threads and they strip.
We can do a lot of things a lot better now days and a lot cheaper to boot but remember BSA had to be competatively priced other wise they would have gone the way of Douglass, Vincient and all the other precision engineered motorcycles.