J-Fall 2023: Roy Wasse & Roy van Rijn
summary
Claim: observable programming skill predicts job performance better than certifications or degrees. A browser-hosted VM (via Apache Guacamole) runs realistic Java/Kotlin tasks that measure coding speed, regression, refactoring and unit test quality while recording edit history; over 250 developers calibrate tasks using peer ratings and the Dunning–Kruger effect; a manual quantitative grading step fixes flaky hidden-tests; results map to junior/medior/senior bands and the platform successfully surfaces surprising overlap between senior and medior scores.
AI-written summary · about
chapters (12)
- 0:00Introduction and speakers
- 1:23Traditional hiring practices
- 4:32Survey insights on developer perception
- 7:07Education and experience impact
- 10:21Research on developer productivity
- 17:01Skill level model and driving analogy
- 24:13Key factors in job performance
- 28:13Measuring programming skill objectively
- 31:42Demo of skill assessment platform
- 43:24Fuj certification initiative
- 47:04Recap and future steps
- 51:53Q&A session
related talks
22mHow good of a developer are you?
53mRefactor your language knowledge portfolio - we are what we can code in
46mThe future of development: Are our jobs getting harder or easier?
01h 45mJava Together… The Power of You - Community Keynote
25mBreak into full stack from Java
46mBeing a Developer: Growing and Developing
32mThe Fake Problem with Junior Developers
01mSteve Yegge, Sourcegraph, ETLS Vegas (10)