03 — Courses That Matter

You are 24, employed full-time, and you can already program. You do not need a course to “learn to code.” You need a course only if (a) your discipline is weak and you want an external forcing function, or (b) a specific instructor’s framing beats the books. Three courses clear that bar in 2026. Everything else on Coursera/edX/Udemy is either redundant with K.N. King or worse.

The three that clear the bar

Course

Provider

Instructor

Cost

Hours

Link

Why this

C Programming with Linux Specialization (7 courses)

edX / Dartmouth + IMT

Petra Bonfert-Taylor et al.

Free audit; ~$540 for full ProfCert

60-90 h

https://engineering.dartmouth.edu/graduate/online/c-programming-with-linux

Best structured “C on Linux from scratch” series. Course 1 (“Getting Started”) is fully free audit. Do C1-C3 for value; C4-C7 optional.

Introduction to Programming in C Specialization (4 courses)

Coursera / Duke

Andrew Hilton, Anne Bracy, Genevieve Lipp

Free audit; ~$49/mo

40-60 h

https://www.coursera.org/specializations/c-programming

Programming-assignment heavy. Duke’s grader will humble you on off-by-one bugs.

CS50x

Harvard / edX

David J. Malan

Free; ~$200 for verified cert

25-40 h (C portion)

https://cs50.harvard.edu/x/

Weeks 1-4 only. Best absolute-first-week onboarding if you want lecture theater energy. Weeks 5+ shift to Python; skip for C purposes.

What most people get wrong about courses

Video courses feel like progress because a bar fills up. Progress is measured in git log, not in “% complete.” If you finish Duke’s assignments without cheating the autograder, that’s real. If you watch every Dartmouth video at 1.5× and skip the exercises, you learned nothing. Rule: one course at a time, and only if you’re doing the exercises.

Return to README.md · Next: 04_youtube_channels.md