VII. Dialogue Preservation
Traditional documentation captures conclusions: “The system uses pattern X.”
It rarely captures the dialogue that led to that conclusion: “We considered patterns X, Y, and Z. Y was rejected because… Z was rejected because… X was chosen because…”
Reasoning Made Visible
Human-AI collaboration sessions, when preserved, capture this dialogue. The reasoning process—including false starts, corrections, and refinements—becomes available to future readers.
This is closer to how theory is transmitted in apprenticeship: not through manuals but through observed problem-solving. The apprentice watches the master think, not just act.
Historical Lineage
This insight has lineage. Nearly a decade before Naur, Kernighan and Plauger recognized that learning happens through concrete examples, not abstract principles:
“Good programming is not learned from generalities, but by seeing how significant programs can be made clean, easy to read, easy to maintain and modify, human-engineered, efficient, and reliable, by the application of common sense and good programming practices. Careful study and imitation of good programs leads to better writing.”
What K&P Couldn’t Preserve
Kernighan and Plauger could only preserve the finished artifact—the good program to be studied and imitated.
The reasoning that made it good, the alternatives considered and rejected, the false starts corrected—these were lost. The reader sees the destination but not the journey.
AI collaboration lets us preserve not just the program but the problem-solving process itself, fulfilling more completely what Kernighan and Plauger could only gesture toward.