A Transient Semantics for Typed Racket

by   Ben Greenman, et al.

Mixed-typed languages enable programmers to link typed and untyped components in various ways. Some offer rich type systems to facilitate the smooth migration of untyped code to the typed world; others merely provide a convenient form of type Dynamic together with a conventional structural type system. Orthogonal to this dimension, Natural systems ensure the integrity of types with a sophisticated contract system, while Transient systems insert simple first-order checks at strategic places within typed code. Furthermore, each method of ensuring type integrity comes with its own blame-assignment strategy. Typed Racket has a rich migratory type system and enforces the types with a Natural semantics. Reticulated Python has a simple structural type system extended with Dynamic and enforces types with a Transient semantics. While Typed Racket satisfies the most stringent gradual-type soundness properties at a significant performance cost, Reticulated Python seems to limit the performance penalty to a tolerable degree and is nevertheless type sound. This comparison raises the question of whether Transient checking is applicable to and beneficial for a rich migratory type system. This paper reports on the surprising difficulties of adapting the Transient semantics of Reticulated Python to the rich migratory type system of Typed Racket. The resulting implementation, Shallow Typed Racket, is faster than the standard Deep Typed Racket but only when the Transient blame assignment strategy is disabled. For language designers, this report provides valuable hints on how to equip an existing compiler to support a Transient semantics. For theoreticians, the negative experience with Transient blame calls for a thorough investigation of this strategy.



There are no comments yet.


page 14


Optimizing and Evaluating Transient Gradual Typing

Gradual typing enables programmers to combine static and dynamic typing ...

Which of My Transient Type Checks Are Not (Almost) Free?

One form of type checking used in gradually typed language is transient ...

On the Use of Underspecified Data-Type Semantics for Type Safety in Low-Level Code

In recent projects on operating-system verification, C and C++ data type...

Shallow Types for Insightful Programs: Grace is Optional, Performance is Not

Languages with explicit dynamic type checking are increasing in populari...

Physical-type correctness in scientific Python

The representation of units and dimensions in informatics systems is bar...

PYInfer: Deep Learning Semantic Type Inference for Python Variables

Python type inference is challenging in practice. Due to its dynamic pro...

CGAL Made More Accessible

We introduce bindings that enable the convenient, efficient, and reliabl...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.