Daisy: An Exercise Environment for Learning Information Modeling
summary
Daisy is a web-based, block-based exercise that breaks information modeling into three concrete steps — information mining, data representation, data example — and supplies natural-language blocks to build data types without syntax. A preliminary classroom study shows most students produce the intended recursive circuit type, but among 31 students who defined the correct type, 15 then build incorrect examples; 6 omit recursion, 8 reuse identical names for constructors and arguments, and 3 duplicate type and argument names. Developers add block-level checks and consider LLM-backed consistency feedback.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:52Information modeling in programming
- 3:22Novice modeling challenges
- 3:42Daisy’s web exercises
- 5:08Information mining in Daisy
- 6:56Data representation in Daisy
- 7:38Data examples in Daisy
- 8:43Daisy effectiveness study
- 10:29Student modeling behaviors
- 13:41Common student mistakes
- 17:35Improving Daisy feedback
- 20:28Conclusion and future work
- 21:15Q&A: Student engagement
- 23:25Q&A: Feedback balance
- 24:30Q&A: Research assumptions
- 25:34Q&A: LLM feedback role
- 27:40Closing remarks
related talks
19mLearn 'em Dafny
07mWeaving Schematics and Code: Interactive Visual Editing for Hardware Description Languages
10mData-centric disambiguation for data transformation with programming-by-example
01h 04mIn Defense of "Little Code"
18mSelf-Directed Learning Material
15mExample-Driven Software Language Engineering
01h 01mDeclarative Programming and Education
01h 39mModeling Data in Haskell for Beginners