Layers Of Complexity

Transcript

1. Layers of complexity Why can't we estimate stuff? Marcus Ahnve @mahnve m@hnve.org
2. About ME • Playing agile coach • Professional developer since 1996 • Polyglot programmer a.k.a web developer • Founder of Agila Sverige
3. Tldr; • Science doesn't work everywhere • Research is not science • People don't know statistics • Software development is not, has never been and will never be a factory
4. Science in software development?
5. Velocity project 1
6. Velocity project 2
7. “We have to get better at DOING time estimates”
8. Square peg, meet round home First we try analysis When that fails we try statistics
9. The problem with science
10. Org anized simplicity Unorganized complexity Medium number systems Complexity Randomness Organized complexity
11. Medium number systems characteristics Anything can happen Murphys Law
12. Sleep Deprivation STUDY
13. We don't really understand Statistics
14. Correlation DOES not imply causation
15. Research Is not Science
16. A/b testing Color # successful # trials % successful Orange 114 4071 2.8 Green 205 6385 3.2 White 59 2264 2.6
17. Binomial Distribution, confidence interval
18. We Really, Really don't understand statistics
19. Science envy
20. Factory envy
21. Frederick W. Taylor 1911 1911 1856 - 1915
22. Economies of Scale
23. Chaos Simple Complex Complicated The cynefin framework
24. We Are doing things for the first time
25. The promise of kanban
26. Scientific method requires repeatability Process 1. Measure output 2. Introduce one change 3. Measure output 4. Evaluate experiment 5. Repeat
27. Standard work Takt time Precise work sequence Standard inventory
28. What is Standard work In software development?
29. Thank you! Marcus Ahnve @mahnve m@hnve.org