Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
167 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

The Essense of Useful Evaluation Through Quantitative Types (Extended Version) (2404.18874v1)

Published 29 Apr 2024 in cs.LO and cs.PL

Abstract: Several evaluation notions for lambda calculus qualify as reasonable cost models according to Slot and van Emde Boas' Invariance Thesis. A notable result achieved by Accattoli and Dal Lago is that leftmost-outermost reduction is reasonable, where the term representation uses sharing and the steps are useful. These results, initially studied in call-by-name, have also been extended to call-by-value. However, the existing formulations of usefulness lack inductive structure, making it challenging in particular to define and reason about type systems on top of the untyped syntax. Additionally, no type-based quantitative interpretations exist for useful evaluation. In this work, we establish the first inductive definition of useful evaluation for open weak call-by-value. This new useful strategy connects to a previous implementation of usefulness through a low-level abstract machine, incurring only in linear time overhead, thus providing a reasonable cost model for open call-by-value implementation. We also propose a semantic interpretation of useful call-by-value using a non-idempotent intersection type system equipped with a notion of tightness. The resulting interpretation is quantitative, i.e. provides exact step-count information for program evaluation. This turns out to be the first semantical interpretation in the literature for a notion of useful evaluation.

Summary

We haven't generated a summary for this paper yet.

X Twitter Logo Streamline Icon: https://streamlinehq.com