I think I know what you are trying to say but you are mixing up characteristics and may confuse members not as smart as you. GoldFinger1969, posted: "The graders would work WITH the computer programmers, just like the auto experts work with them. A bunch of coin graders would tell for each particular coin what they look for and the programmers would then write the code that would tell the computer what to look for. It would be "smart" enough to know how to scan for luster vs. reflectivity....[But luster is the reflectivity of light from the surface] to distinguish bag marks from polishing marks....[these look nothing alike and result from different causes] to look for rub vs. deep coin gouges [ditto] from the bag." Furthermore, how are you going to program the computer to ignore friction wear like the TPGS's have done for decades. LOL.
Great Discussion I have been involved with many precision components. I know that it would be pretty straightforward to digitize a coin's surface and comparing it to either another coin or even the die. It would be fairly easy to evaluate the surface for strike, wear, and detail. Finish and luster can also be determined numerically, but the location of the readings could be subjective. "Eye Appeal" is a big fly in the ointment. A nice design has more appeal than a lousy design. "Eye Appeal," in my mind, is in the eye of the beholder and only relates to $$.
Potentially, it will take time to use the computer such as load the coin or get it scanned etc. The time aspect will probably be a push
True computer grading will be very simple when it is finally done decades from now. There will be no need to position a coin and they can be fed into the "box" on a belt with clear holes so each side will be examined at once. The computer will ignore any smudges or scratches on the glass and after the coin is graded and removed it could even be run through a cleaning station before cycling back to the robot that is placing the coin on the belt. LOL.
At which point it will be unnecessary as robot collectors could just grade it themselves off the same program That was only half tongue in cheek, but even in the future a completely robot run system will be prohibitively expensive and unlikely to actually come to fruition for coin grading. If we're at that point where robots are doing everything no one will have any money anyways
Heck, just flip the coin across a gap surrounded by bright lights and high-speed cameras, then reconstruct the coin's surface detail by analyzing the video streams. The resulting model will be more detailed than any human's view of the coin could ever be. At that point, the coin itself is irrelevant; you can just let it land in a melting pot.