Quit Simplifying!
summary
Argues that software complexity is inevitable and measurable via the configuration complexity clock (noon → config → rules → DSL → regenerate), with concrete cases: Ansible’s easy-to-start-but-grow history, Open edX locking in mongodb via GridFS a decade ago and failing to remove it because a Blockstore refactor hit severe performance problems, and Tuda generating Kubernetes and Docker manifests so plugins push 10–15x complexity onto admins. Recommends concrete controls: always ship docs, tests, and a feature flag with features, deprecate before refactor, and treat managers who add channels as entropy accelerators.
AI-written summary · about
chapters (16)
- 0:00Welcome and introduction
- 1:32Ansible's simplicity claim
- 4:02New systems feel simpler
- 6:07Configuration complexity clock
- 9:57Ansible's complexity journey
- 12:46Open edX complexity example
- 15:36Complexity permanence in projects
- 20:57Open edX deployment complexity
- 23:07Tutor and Docker simplification
- 26:58Pushing complexity downstream
- 29:04Physics of complexity
- 33:23Drivers of software entropy
- 35:11Slowing disorder growth
- 37:37Managerial entropy acceleration
- 40:22Keep things simple
- 41:26Q&A
