Breaking it down, there are basically two components to ballistic fingerprinting; one method is by the rifling marks left on the bullet; the other is by unique markings left on
the cartridge case. AFAIK, right now the only component being logged right now is the casing.
Not that it really matters because as you pointed out, the characteristics of both components change after time through use, and can be deliberately altered. A barrel, firing pin or bolt change. Hand and fire lapping. A file and stone to the bolt face and firing pin, etc.
Glocks with hexagonal rifling are very difficult to match. With human fingerprints no two are alike, even with twins. If a person tries to alter them they simply regenerate. How precise is firearms ballistics? If a number of handguns were made on the same day, with the same tools and machines by the same operators how much variation is there between the first and second
firearm off the line? How much difference between the first and one-hundredth?
Even if it were a foolproof science, ballistics finger printing is back door registration. A ballistic database does no good if there is no owners name to go with the firearm.
Is anyone proposing a national database of tire marks or shoe prints? How about every citizen submitting a sample of their handwriting?
I'm not sure if this is an urban legend or not but, back in the days when people used typewriters, typewriters in the Soviet Union were registered and a typewritten sample was kept on file so the authorites could trace the identity of a letters author.