In my experience, the methods used in ink key presetting systems are based on the same principle used intuitively by press operators. I.e. eyeball the ink coverage in each ink zone and estimate the amount the keys need to be opened in order to deliver the needed amount of ink.
In that context ink key presetting systems are much faster and more accurate than the press operator. Some solutions can also "learn" by comparing what was predicted vs what was actually needed and modifying the algorithms accordingly.
And in that sense it is capable and it doesn't matter if the people who designed the system understand the problem or not.
The system may not be optimal, but it appears to be sufficiently better than eyeballing the plate. At least until a better method is developed.
best, gordo
I would ask you. Do you understand the problem? If you don't then you would not know what was wrong. You would also not know that self learning algorithms are not workable with this type of problem because for each new job, the problem to solve is different. The idea of iterative adjustments does not work even though it might appear to have some improvement. One could do a series of jobs as a test that would totally screw up any self learning algorithm.
The solution to the presetting of ink keys is not directly related to the image coverage area, so even if the computer accurately calculates the image area instead of the eyeball approach, it only means that the calculation is a more precise calculation of the wrong data.
I am not saying that an incorrect algorithm is useless. Even poor algorithms are better than what the operator can do. That is not the point. If one does not care about accuracy then I would agree with you that what exists is sufficient, but unfortunately I do think accuracy and better technical solutions are important.
I am a little surprised at your satisfaction in aiming low. The calculation for ink key presetting does not requires high precision but it does need to be accurate.
The presetting capability going from eyeballing the plate to measuring the image area has been in use for about 30 years. Presetting is not a new technology and one should expect that the knowledge would have advanced. Knowledge does not advance when people are satisfied with mediocre performance. When satisfied, there is never the need to be curious about what would work better.
So if one is happy to have a press where expensive closed loop is needed or where the operator needs to make adjustments that results in extra paper being printed instead of the job going directly to the target densities, it is hard to argue with that reality.
Gordon, let me add an additional comment.
I would guess that the engineers at Creo that you knew, would not be happy or satisfied if they thought an algorithm was not good enough. Users of an improved technology might not think that is an issue but for those who design the technology, knowing that it could be better will bother them. It is the knowing that it could have been better that will drive a designer crazy. It can be viewed as a lost opportunity instead of the success that the market might see it.
/ is little bit coarse ... weak price/performance ratio for today's challenging times. Just my opinion (but based on real experiences).
