In this informal talk I will tackle with the problem of logical errors detection in program's construction. It common to encountered logical errors when we are building software. The main difference between these errors and other errors is that they cannot be detected by any debugger, parser or other automatic/semi-automatic tool. That is because they are syntactically correct, but their existence will become pernicious for the program's behaviour and the user will have to face undesirable side effects: wrong results,infinite loops, missing answers, failures, etc. I will propose a system, at a theoretical level, that will be able, to some extend, of 'trapping' some of these errors. The system will be based on the fact that the operation of the program's construction will be monitored by a techniques editing system which will convey to the system some of its strong characteristics. Possible solutions for different approaches to the problem as well as ways of controlling the program's construction will be discussed. Your input -comments/suggestions- will be welcomed.