Doing I.T. for SCIENCE! - Sprints, Startups and the Scientific Method
summary
Argues for treating feature work as experiments: observe abandoned carts, formulate hypothesis, run A/B testing to split traffic, analyze weeks of logs, and discover a Firefox payment failure caused by an EV certificate; isolates a bug where NHibernate sent Unicode strings breaking non-Unicode indexes; recommends building throwaway prototypes with Axure, shipping a MVP/MLP, running short two-day investigative spikes, and publishing results so future teams reuse findings instead of repeating the same blind work.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 1:04Computing and Science
- 4:06Word Processor
- 6:29On Start Again
- 7:44Super Mario
- 8:58Reusing Existing Solutions
- 9:29Cutting Edge Solutions
- 10:38The Scientific Method
- 13:44Rainbows
- 14:21Newton
- 15:13Newton Hypothesis
- 17:41Newtons Vision
- 18:58Observation
- 24:24Experiment vs Product
- 27:10The Pilot System
- 29:04The Illusion of Choice
- 32:16Agile Product Development
- 33:43Scientific Experiment
- 36:32Dependency Injection
- 38:29Easy Code Reuse
- 41:19Experimental Validation
- 42:01Analysis
- 42:20Better
- 42:49Facetoface
- 43:46Questions
- 46:02Conclusion
related talks
24mEffective Product Experimentation: Using Data-Driven Techniques for Iteration
35mExperimentation for Engineers
21mNeed for speed - Moving fast as a software engineer
36m"Engineering to embrace change" Or "How clumsy engineering can lead to powerful research – Part II"
57mThings That Work So Well for Me I Can't Believe You're Not Using Them
48mExperiments, the Good, the Bad, and the Beautiful
46mCrescent Wrenches and Debuggers
49mHow to get smarter while doing your job