01 / The problem
When the software gets in the way of learning.
Beginning programmers need a manageable place to write, run, trace, and test code. A separate challenge for CS50P students is getting started without being handed a solution. These projects address two related needs with tools designed for student independence.
02 / Learning products
Two tools, complementary roles.
Python Lab
An accessible coding workspace that runs Python in the browser with an editor, locally saved projects, starter activities, a simulated terminal, testing support, and a variable-by-variable trace table.
Try Python Lab ↗CS50 Learning Companion
A solution-free VS Code extension offering 41 assignment guides, incremental hints, targeted Python references, testing shortcuts, and links to Harvard's official problem requirements.
View extension ↗03 / Instructional judgment
Scaffold the next action, not the answer.
- Progressive support: CS50 Companion uses four levels of starter scaffolding across the problem sets without completing assignment solutions.
- Prediction before execution: Python Lab's Debug Trace guides students to predict code behavior, then inspect variable changes in a step table.
- Reduce interface overload: nonessential tabs and controls were consolidated based on student feedback about complexity.
- Preserve learner work: existing Python files are never overwritten by Companion updates. Python Lab stores work in the browser profile and offers explicit project export and backup.
- Use authentic specifications: Companion references official assignment instructions and preserves required file, function, and test structures.
04 / Prototype & validate
Use real workflows to improve the product.
Both projects grew through iterative classroom use: identifying friction, creating smaller testable improvements, validating behavior with automated checks, and reviewing what students actually see. The repositories document implementation, tests, and changes over time.
05 / Evidence boundary
Observable features, not invented learning gains.
Implemented editor, trace, scaffolding, and testing workflows are documented in the current public repositories. The projects are working examples of self-directed learning support and rapid product iteration.
Evidence boundary: No verified improvement in assignment completion, mastery, or learning time is claimed here. User feedback and test coverage demonstrate an iterative process, not proof of student learning effectiveness.
Let's make learning useful.
Let's talk about building effective learning experiences for teams and organizations.
Contact Andrew ↗