Things we lost in the fire
summary
Multiple client projects show what to watch for: entrenched teams that reject sensible tech (XSLT resisted by devs), tools with hard limits that break design (editor attribute cap 8,000 chars, first login 30 minutes), and process failures that waste months (45-person standup, six-week waterfall schema hire). Migration pain appears in scale—tens of thousands of documents, 20–25 languages—and in politics where workarounds win (DTD converted with Klingon attributes) or vendors force choices that fail (SGML→XML rewrite leads to lawsuit). Walk away if licensing, tooling limits, or culture block real change.
AI-written summary · about
chapters (14)
- 0:00Introduction and setup
- 1:27Grumpy old man origins
- 2:36XSLT religious war story
- 5:10Automotive manuals disaster
- 8:02Waterfall project absurdity
- 9:51Legacy DTD and Swedish tags
- 13:25Ten-computer bottleneck
- 15:59Open-source policy madness
- 16:19Car manufacturer XML failure
- 20:39SGML vs XML aerospace case
- 22:30War stories and lessons
- 24:17Engagement and battles
- 26:53Contract negotiation tips
- 27:36Q&A and closing
related talks
47mLegacy Software: Really a Problem?
59mArchitecture Modernization Anti-Patterns: The Cost of Good Intentions
44mLegacy Software Really a Problem
40mLegacy Evolution – The Innovation Opportunity!
43mHow to Innovate in a Legacy
44mCruft and Technical Debt: A Long View
54mCtrl-Alt-Del: learning to love legacy code
19mHalt and Don’t Catch Fire