Today even a fairly typical medium-size software system often runs to one hundred thousand lines of code. Large systems often have roughly a million lines of code. Very large systems have ten million lines of code, and the largest systems in operation have roughly one hundred million lines of code.
No human can even come close to comprehending such systems as a unit.
One of the best programmers I've worked with, Hugh Holbrook, puts the limit at about ten thousand lines of code, and organizes his work in terms of loading the most relevant ten thousand lines of code into human working memory, completing the task, and then moving on to repeat the process on the next task.
This leads us to the first fundamental principle of comprehensible computing:
No large software system is humanly understandable as a whole. To be comprehensible at all, large software systems must be piecewise-understandable in mindquanta of ten kilolines of code or less.
Back to Comprehensible Computing.