Source Reader

Literature is taught by reading literature. Code is taught by writing it. This has always been a mistake.

Source Reader is an anthology of masterpieces of software. Not the most technically complex. The ones where knowing who wrote them, and when, and under what pressure, changes what you see when you read them. Code is not produced in a vacuum. It carries the decisions, constraints, and blind spots of the moment it was made. Reading it seriously means reading all of that.

Each entry is a close reading. A person, a moment, a piece of code that still runs — or that shaped what runs today. The model is the museum catalogue essay, not the CS textbook. The artifact is always the center. Everything else is context.

The urgency is structural. AI systems now generate syntactically correct code on demand. The scarce skill is no longer writing code — it is judging it. Knowing whether it is right, elegant, appropriate, or dangerous requires exposure to work that set the standard and work that failed to.