Teaching Python like a discovery, not a syllabus

The fastest way to kill curiosity about programming is to explain everything first. We hand students the syntax, the theory and the "best practice" before they have ever felt the need for any of it — and then wonder why the knowledge doesn't stick.

A peer-driven, project-first model turns that order around. Students are given a problem slightly beyond what they know, minimal scaffolding, and each other. They struggle, they search, they ask the person next to them, and the concept arrives exactly when they have a use for it.

What changes when the lecture stops

• The question comes before the answer. A student who has already hit the wall remembers the fix. A student who was warned about the wall does not.
• Peers become the first line of help. Explaining to someone one step behind you is where understanding actually consolidates.
• The lecturer becomes an editor, not a broadcaster — pointing, unblocking, raising the bar.

It is not chaos

Project-first learning only works with real structure underneath: a tight sequence of problems, clear checkpoints, and honest assessment of what was actually built. The freedom is in how a student gets there, not in whether they do.

I'm redesigning my introductory and graduate computing courses around this idea. Less telling, more discovering — and far more that survives past the exam.