Haninah Levine has the second of what will be four articles at DefenseTech.org on the Reliable Replacement Warhead program.
In these cases, a dilemma arises: should the nuclear production complex go to extreme lengths to recreate the processes needed to remanufacture [nuclear weapons] components exactly according to the original specifications? Or should they look for ways to make replacement parts that will work just as well, if not better? Since the part has to be replaced anyway, why not make maintenance easier for future generations already?
[snip]
If too many of these minor changes pile up, though, a sort of “Grandfather’s axe” effect may kick in: if enough components have been modified and replaced, is the warhead design still the same one that was once tested? For this reason, the guiding philosophy has been “change-control discipline”: make the fewest number of changes possible, and only after proving exhaustively that the changes will not affect warhead characteristics.