The hard road from methods to practice [software engineering]

Software engineering methods-by which the author means any strategic or tactical ideas intended to guide the practice of developing software-are tools. A methodology gap exists whenever a stated method doesn't correspond to effective practice. A methodology gap can manifest in several ways. It could be that a method can't be practised, shouldn't be practised, or simply isn't practised as designed. The gap is not necessarily a bad thing-we can productively strive for ideals that will never be perfectly achieved. But when such a gap goes unrecognized, or calcifies into a tradition of self deception, it can suck the marrow out of your efforts to control or predict your software development process. The causes of a methodology gap are usually subtle and complex. Methodologists or management may not understand the realities faced by practitioners. Practitioners may not understand or respect the methods. Or they may forget them accidentally or apply them recklessly. The methods may be poorly designed or poorly communicated. The gap can also perpetuate itself when practitioners know that publicly accepted methods aren't really followed; what incentive do they have to embrace any new idea that comes along? Process hypocrisy leads to process cynicism and vice versa. A vicious cycle.