Computer Science, Software Engineering, and the Scientific Method
summary
Argues that software sits between engineering, science and art and that concrete improvements come from using scientific practice: double down on testing, isolate changes and run statistical benchmarks. Cites that NASA cuts defects to near zero only by massive manual review and order-of-magnitude cost increases, so the goal is cheaper rigour. Recommends tooling—unit tests, benchmarking, logging, Jenkins/GitHub CI, Python introspection and TDD—plus teaching software engineering as craft to improve estimates, composability and reproducible behaviour.
AI-written summary · about
chapters (29)
- 0:00Introduction
- 0:21Rackspace
- 0:34Python Software Foundation
- 1:09We are not developers
- 1:36A show of hands
- 2:18Three types of members
- 3:05Our goal
- 3:26Our work
- 4:09PyCon 2014
- 5:08Programming is a young profession
- 5:40Software is bugs
- 5:54Computers are terrible
- 6:51Moores Law
- 7:16Software Delivery
- 10:10Applied Science
- 11:40Science
- 13:32Epistemology
- 15:02Isolation
- 16:16Observing
- 16:53Testing
- 19:23Testdriven development
- 20:11Programming is like other industries
- 21:17Teaching programming
- 21:47Software engineering curriculum
- 23:57Please help me
- 24:27Chris Bosh
- 26:37Alex Hilder
- 35:19Tooling
- 37:49Design
related talks
45mTaking Back "Software Engineering"
42mwhy software development is an engineering discipline
35mCraft, Engineering, and the Essence of Programming- 42mwhy software development is an engineering discipline
32mSoftware Craftsmanship
42mWhat our Brains can Teach us About Software Engineering
48mEngineering You
44mBehavioral Software Engineering