In a previous article on constraint discovery in COSA, I wrote about its inductive nature. Inductive means that discovered constraints are not written in stone. It is up to the program designer to either validate or reject every discovered constraint. COSA features another type of constraint, a categorical one. It is governed by the principle of motor coordination or PMC. This principle will eliminate every kind of internal logical conflict (inconsistency) within a COSA program. I call it categorical because every discovered PMC conflict must be eliminated. Note that PMC conflict discovery can also be completely automated. In conclusion, the combination of automatic constraint discovery (inductive and categorical) and the automatic resolution of dependencies solve the reliability problem in COSA programs. In a future page, I will explain the PMC in greater detail.
Note. Constraint discovery is part of my ongoing work in artificial intelligence. There is a way to combine inductive and categorical constraint discovery into a single mechanism that will enable its full automation. Unfortunately, I cannot divulge this method at this time. Sorry.