Harry Potter and The Legacy Code Base
summary
Claim: legacy code is a storehouse of domain knowledge, fixable by archaeology not heroics. Start with surveillance, excavation, analysis — gather metrics (test coverage, New Relic), hunt for red flags (flog score > 1,000), then perform targeted small changes: rename and extract (e.g. remove_poison → harvest_poison, PoisonBarrel class), deploy fresh hires to explore, keep tests green, document with UML, READMEs and wikis. Four students were brought in as examples; incremental, measured work proves effective.
AI-written summary · about
chapters (11)
- 0:00Hogwarts legacy code battle
- 2:05Legacy code beasts in basement
- 5:12Legacy code as knowledge store
- 9:06Code archaeology 101
- 14:12Be Neville Longbottom of legacy code
- 15:14Surveillance first technique
- 19:30Excavation and refactoring plan
- 24:54Analysis and documentation
- 29:05Legacy code beasts survival
- 31:09Breaking legacy code phantasms
- 31:34Q&A
related talks
28mHarry Potter and the Legacy Codebase
23mLoving Legacy Code
39mDown The Rabbit Hole: An Adventure in Legacy Code
01h 25mGetting a Handle on Legacy Code
38mTackling Large Ruby Refactorings with Confidence (and Science!)
36mWorking Compassionately with Legacy Code
34mTesting legacy code when you dislike tests (and legacy code)
51mCode spelunking in the All New Basecamp