A crucial aspect of design is knowing which things need to be part of a module and which can be (or should be) part of separate modules. The guiding mantra is "Systems should be built from cohesive, loosely coupled components (modules)." Some aspects of a system have numerous dependencies, and dividing those aspects causes more harm than good. Others have few to no dependencies, and tossing them into a single module makes it needlessly complex. Such modules make it challenging to compellingly design the disparate items and confusing to those who follow and need to update the system. "Divide and Conquer" is seldom successfully done with a sword, as in the parable of the Judgement of Solomon.
The essence of deconstructing an unknown is working to understand the intent and the design without destroying it. (E.g., don't use a pry bar when all that is needed is a screwdriver.) Without all of the effective and maintained key work products needed to create and sustain a system, it is usually less costly in time and money to dispose of it and start anew.
As long as it remains as a black box, it is highly unwise to think you can make any beneficial change to it.
I did not say that swords have no useful purpose and I did say "seldom". My point is that knowing when to use a sword requires insights into the positive and negative consequences.