How we build lessons
Every lesson goes through the same pipeline before it is published.
The pipeline
1Specification
A human writes what the lesson must teach: the goal, the concepts that cannot be skipped, the kind of practice, and what is explicitly out of scope.
2Draft
The draft is produced against that specification. Its structure is machine-checked: a lesson missing theory blocks, practice or hints does not pass.
3Validation
Automatic checks verify the structure, the prerequisite graph and that no text is hardcoded outside the translation files.
4Technical review
Claims are checked for factual accuracy. Simplifications are allowed; statements that are simply false are not.
5Translation
Each lesson is translated with a fixed glossary, so terminology stays consistent across the whole course.
6Publish
The lesson ships only when every check passes. If the checks fail, the build fails — broken content cannot reach the site.
What is checked automatically
Every lesson must parse against a schema, every prerequisite must exist, every task must have at least two hints, every user-visible string must exist in all supported languages, and every simulator value must match the text that describes it.
Limits, stated honestly
Lessons are written with machine assistance and then reviewed. That means errors are possible, and we would rather say so than pretend otherwise. If you find one, report it — corrections are applied to the lesson, not hidden.