Software Delivery and the Rube Goldberg Machine: What Is the Problem We Are Trying to Solve?
summary
Compares delivery pipelines to Rube Goldberg machines and names concrete pain points: dozens of brittle tools, fragile glue code, dependency surprises like GitHub changing archive checksums, and a Guinness-record machine of 427 steps that takes 4m26s as a cautionary metaphor. Cites tool examples and efforts: JFrog, salsa.dev checklists, CNCF/CD Foundation work and the emerging CD Events spec for interoperability. Notes measurable risks—76% of observed containers run as root across ~3 million containers—and practical fixes.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:47Rube Goldberg vs software
- 5:17Rube Goldberg traits
- 8:04Software delivery ingredients
- 12:33CNCF landscape
- 14:00Documenting tool choices
- 16:28Tool misuse risks
- 18:57Pipeline communication gaps
- 20:41Interoperability challenges
- 21:42Scope creep risks
- 22:27Pipeline reliability
- 23:32Supply chain security
- 25:17Base image risks
- 28:31Industry solutions
- 29:38CD events standards
- 30:43Closing & Q&A
related talks
26mSoftware Delivery from First Principles
50mWhy You Need a Software Delivery Machine
41mContinuous delivery in 10 easy mistakes
30mBringing Continuous Delivery to Open Source
55mContinuously Delivering Continuous Delivery
49mModern Software Delivery: Triumphs and Failures
26mWhere Are We Going? The Future of Continuous Delivery
23mDay 1 Welcome Keynote - Simplifying Software Delivery