Attractive as it seems at first blush, the idea of a database of "ballistic fingerprints" is not as good as might seem. With use, the characteristics of a barrel (as to the projectile markings) or the chamber (as to the shell markings) will change to some degree, making a "true match" less likely and a "near match" of less value for proof. This is also true for typewriter-to-typewritten-sample matching, which in ideal circumstances is comparable in value to ballistics testing.