Several studies have been done on the viability of micro-serial engraving, barcoding, or stamping on various firearm parts.
One of my favorites was done by some folks at the University of California. (The study was performed in favor of micro-serial engraving, but showed that it's a dead-end with most designs.)
The Study.
The Appendices. (The study isn't as useful without the appendices.)
One of the most important statements in there; is that in a blind test, the micro-stamped shell casings had an average of 3 completely different numerical sequences identified by the subjects for each casing. In nearly all of those cases, those sequences differed from the sequences identified by the study's authors, as well. So, each, individual casing had up to 5 different numerical sequences identified from the impression that was left... and most were completely wrong.