Learning to program is hard, and how to fix that
summary
Course material gets redesigned to force and surface real mistakes: a my pie tutor system runs static analysis and returns targeted hints, tutorial problems include an is_dna validator that exposes off‑by‑one and boolean-return errors, and a debugging exercise with six planted bugs trains iterative fixes. Data points include almost 1,000 students tutored and class size rising past 600, and measured outcomes show some tasks still take up to half an hour so pacing and error messages get adjusted.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:45Who am I
- 1:30What does a tutor do
- 2:50Class size
- 3:17Learning Environment
- 5:22Language Documentation
- 6:50Brain Scanner
- 7:23Problem Solving Skills
- 8:17First Task
- 17:08Constraints
- 18:27Tasks
- 20:46Summary
- 22:11Python error messages
- 23:01Traceback messages
- 24:10Types
- 25:48Static analysis
- 26:38Debugging
- 28:00Too much help
related talks
30mTeaching Programming for the fun of IT
27mLearn Python Like a 12 Year Old
23mLearning to learn Python
36mTrials & tribulations teaching Python at CoderDojo
31mTeaching Python Online to Primary Students
29mAssurance of Learning Through Testing
26mTeaching teachers to teach Python: what have I learned?
30mConcept to Classroom: Creating fun differentiated programming projects