Keynote - THE MISMEASURE OF SOFTWARE
summary
Claims that many common metrics are wrong and harmful, with two concrete failure modes: the map‑territory error (model ≠ reality) and reification (counting immaterial things). Specific substitutes and tools discussed include defect density, defect detection percentage, lines of code and velocity, plus the antique psychograph as a cautionary build that convinces but fails. Practical rules are the four don'ts (know the meaning, act on it, measure accomplishment not effort, avoid targets), and the favorite signal is surprise as an indicator with intake/meaning/significance diagnostics.
AI-written summary · about
chapters (21)
- 0:00Start
- 4:41A Classic Example
- 5:12A Volunteer From The Audience?
- 9:54Two Causes of Measurement Failure
- 11:03Map-Territory Error
- 14:59What Is Measurement?
- 18:02Reification Error
- 23:25The Four Don'ts
- 25:02Don't #1
- 27:39Don't #2
- 27:58Don't #3
- 29:59Don't #4
- 34:31Topics
- 37:51A Better Set
- 39:57My Favorite Indicator Is
- 42:22Satir Interaction Model
- 45:33Intake Surprises
- 47:02Meaning Surprises
- 49:08Significance Surprises
- 52:03Lee's Zeroth Law of Measurement
- 53:17Recap
related talks
31mMetrics Mania. Obsession with Measurement can Harm our Teams
35m"You Can't Miss What You Can't Measure"
06mDelivery Metrics: the good, the bad, and the totally ridiculous
29mYour Metrics Are Lying: How to Stop Reporting and Start Learning
39mTest Metrics – a thread or a threat
18mFrom developer productivity to engineering impact
47mCan We Measure Developer Productivity?
01h 01mKeynote: Build Good Software: Of Politics and Methods