Schemas for the Real World
summary
Argues that schemas must mirror messy human identity rather than force reductive lists, showing that turning gender into a text field for gender (MetaFilter, Diaspora) yields richer, honest profiles while rigid options produce misleading, coerced answers; gives concrete patterns—use auto-suggest with overrides (LinkedIn) or allow nulls early, model relationships as one-to-many relationship modeling not a single status—and cites ARIS finding roughly 100 unique religion answers later compressed for analysis, so start loose and refine.
AI-written summary · about
chapters (8)
- 0:00The human cost of rigid data
- 1:40Flawed premises in design
- 6:03Realigning mental schemas
- 9:24Learning from open text fields
- 11:54Internationalization challenges
- 18:14Data quality vs. reductionism
- 25:32Building for the real world
- 30:12Q&A and closing thoughts
