This article presents an algorithmic reconstruction of Normalisation by Evaluation (NbE), exploring why this technique outperforms substitution-based normalisers. The author demonstrates how to identify inefficiencies in naive approaches and optimize them to achieve NbE, ultimately developing a normaliser that is asymptotically superior to standard NbE implementations in some cases.
Background
Normalisation by Evaluation is a technique for converting programs to their normal forms by first evaluating them in a semantic domain and then reifying the result. It's known to be more efficient than substitution-based approaches but the reasons for its performance advantages haven't been fully explained.
- Source
- Lobsters
- Published
- Apr 23, 2026 at 01:17 AM
- Score
- 7.0 / 10