Polymorphic Lambda Calculus Research Articles

Parallel Algebraic Effect Handlers

Algebraic effect handlers support composable and structured control-flow abstraction. However, existing designs of algebraic effects often require effects to be executed sequentially. This paper studies parallel algebraic effect handlers. In particular, we formalize λ p , a lambda calculus which models two key features, effect handlers and parallelizable computations, the latter of which takes the form of a for expression, inspired by the Dex programming language. We present various interesting examples expressible in our calculus. To show that our design can be implemented in a type-safe way, we present a higher-order polymorphic lambda calculus F p that extends λ p with a lightweight value dependent type system, and prove that F p preserves the semantics of λ p and enjoys syntactic type soundness. Lastly, we provide an implementation of the language design as a Haskell library, which mirrors both λ p and F p and reveals new connections to free applicative functors. All examples presented can be encoded in the Haskell implementation. We believe this paper is the first to study the combination of user-defined effect handlers and parallel computations, and it is our hope that it provides a basis for future designs and implementations of parallel algebraic effect handlers.

Polymorphic lambda calculus with context-free session types

Session types provide a typing discipline for structured communication on bidirectional channels. Context-free session types overcome the restriction to tail recursive protocols characteristic of conventional session types. This extension enables the serialization and deserialization of tree structures in a fully type-safe manner.We present the theory underlying the language FreeST 2, which features context-free session types in an extension of System F with linear types and a kinding system to distinguish message types, session types, and channel types. The system presents metatheoretical challenges which we address: contractivity in the presence of polymorphism, a non-trivial equational theory on types, and decidability of type equivalence. We also establish standard results on typing preservation, progress, and a characterization of erroneous processes.

Finitary polymorphism for optimizing type-directed compilation

We develop a type-theoretical method for optimizing type directed compilation of polymorphic languages, implement the method in SML#, which is a full-scale compiler of Standard ML extended with several advanced features that require type-passing operational semantics, and report its effectiveness through performance evaluation. For this purpose, we first define a predicative second-order lambda calculus with finitary polymorphism, where each type abstraction is explicitly constrained to a finite type universe, and establishes the type soundness with respect to a type-passing operational semantics. Different from a calculus with stratified type universes, type universes of the calculus are terms that represent a finite set of instance types. We then develop a universe reconstruction algorithm that takes a term of the standard second-order lambda calculus, checks if the term is typable with finitary polymorphism, and, if typable, constructs a term in the calculus of finitary polymorphism. Based on these results, we present a type-based optimization method for polymorphic functions. Since our formalism is based on the second-order lambda calculus, it can be used to optimize various polymorphic languages. We implement the optimization method for native (tag-free) data representation and record polymorphism, and evaluate its effectiveness through benchmarks. The evaluation shows that 83.79% of type passing abstractions are eliminated, and achieves the average of 15.28% speed-up of compiled code.

Modules, abstraction, and parametric polymorphism

Reynolds's Abstraction theorem forms the mathematical foundation for data abstraction. His setting was the polymorphic lambda calculus. Today, many modern languages, such as the ML family, employ rich module systems designed to give more expressive support for data abstraction than the polymorphic lambda calculus, but analogues of the Abstraction theorem for such module systems have lagged far behind. We give an account of the Abstraction theorem for a modern module calculus supporting generative and applicative functors, higher-order functors, sealing, and translucent signatures. The main issues to be overcome are: (1) the fact that modules combine both types and terms, so they must be treated as both simultaneously, (2) the effect discipline that models the distinction between transparent and opaque modules, and (3) a very rich language of type constructors supporting singleton kinds. We define logical equivalence for modules and show that it coincides with contextual equivalence. This substantiates the folk theorem that modules are good for data abstraction. All our proofs are formalized in Coq.

System f-omega with equirecursive types for datatype-generic programming

Traversing an algebraic datatype by hand requires boilerplate code which duplicates the structure of the datatype. Datatype-generic programming (DGP) aims to eliminate such boilerplate code by decomposing algebraic datatypes into type constructor applications from which generic traversals can be synthesized. However, different traversals require different decompositions, which yield isomorphic but unequal types. This hinders the interoperability of different DGP techniques. In this paper, we propose Fωμ, an extension of the higher-order polymorphic lambda calculus Fω with records, variants, and equirecursive types. We prove the soundness of the type system, and show that type checking for first-order recursive types is decidable with a practical type checking algorithm. In our soundness proof we define type equality by interpreting types as infinitary λ-terms (in particular, Berarducci-trees). To decide type equality we β-normalize types, and then use an extension of equivalence checking for usual equirecursive types. Thanks to equirecursive types, new decompositions for a datatype can be added modularly and still interoperate with each other, allowing multiple DGP techniques to work together. We sketch how generic traversals can be synthesized, and apply these components to some examples. Since the set of datatype decomposition becomes extensible, System Fωμ enables using DGP techniques incrementally, instead of planning for them upfront or doing invasive refactoring.

Dependent types and multi-monadic effects in F*

We present a new, completely redesigned, version of F*, a language that works both as a proof assistant as well as a general-purpose, verification-oriented, effectful programming language. In support of these complementary roles, F* is a dependently typed, higher-order, call-by-value language with _primitive_ effects including state, exceptions, divergence and IO. Although primitive, programmers choose the granularity at which to specify effects by equipping each effect with a monadic, predicate transformer semantics. F* uses this to efficiently compute weakest preconditions and discharges the resulting proof obligations using a combination of SMT solving and manual proofs. Isolated from the effects, the core of F* is a language of pure functions used to write specifications and proof terms---its consistency is maintained by a semantic termination check based on a well-founded order. We evaluate our design on more than 55,000 lines of F* we have authored in the last year, focusing on three main case studies. Showcasing its use as a general-purpose programming language, F* is programmed (but not verified) in F*, and bootstraps in both OCaml and F#. Our experience confirms F*'s pay-as-you-go cost model: writing idiomatic ML-like code with no finer specifications imposes no user burden. As a verification-oriented language, our most significant evaluation of F* is in verifying several key modules in an implementation of the TLS-1.2 protocol standard. For the modules we considered, we are able to prove more properties, with fewer annotations using F* than in a prior verified implementation of TLS-1.2. Finally, as a proof assistant, we discuss our use of F* in mechanizing the metatheory of a range of lambda calculi, starting from the simply typed lambda calculus to System F-omega and even micro-F*, a sizeable fragment of F* itself---these proofs make essential use of F*'s flexible combination of SMT automation and constructive proofs, enabling a tactic-free style of programming and proving at a relatively large scale.

Efficiency of lambda-encodings in total type theory

AbstractThis paper proposes a new typed lambda-encoding for inductive types which, for Peano numerals, has the expected time complexities for basic operations like addition and multiplication, has a constant-time predecessor function, and requires only quadratic space to encode a numeral. This improves on the exponential space required by the Parigot encoding. Like the Parigot encoding, the new encoding is typable in System F-omega plus positive-recursive type definitions, a total type theory. The new encoding is compared with previous ones through a significant case study: mergesort using Braun trees. The practical runtime efficiency of the new encoding, and the Church and Parigot encodings, are compared by two translations, one to Racket and one to Haskell, on a small suite of benchmarks.

Wellfounded recursion with copatterns

In this paper, we study strong normalization of a core language based on System F-omega which supports programming with finite and infinite structures. Building on our prior work, finite data such as finite lists and trees are defined via constructors and manipulated via pattern matching, while infinite data such as streams and infinite trees is defined by observations and synthesized via copattern matching. In this work, we take a type-based approach to strong normalization by tracking size information about finite and infinite data in the type. This guarantees compositionality. More importantly, the duality of pattern and copatterns provide a unifying semantic concept which allows us for the first time to elegantly and uniformly support both well-founded induction and coinduction by mere rewriting. The strong normalization proof is structured around Girard's reducibility candidates. As such our system allows for non-determinism and does not rely on coverage. Since System F-omega is general enough that it can be the target of compilation for the Calculus of Constructions, this work is a significant step towards representing observation-centric infinite data in proof assistants such as Coq and Agda.

Mixin’ Up the ML Module System

ML modules provide hierarchical namespace management, as well as fine-grained control over the propagation of type information, but they do not allow modules to be broken up into mutually recursive, separately compilable components. Mixin modules facilitate recursive linking of separately compiled components, but they are not hierarchically composable and typically do not support type abstraction. We synthesize the complementary advantages of these two mechanisms in a novel module system design we call MixML.A MixML module is like an ML structure in which some of the components are specified but not defined. In other words, it unifies the ML structure and signature languages into one. MixML seamlessly integrates hierarchical composition, translucent ML-style data abstraction, and mixin-style recursive linking. Moreover, the design of MixML is clean and minimalist; it emphasizes how all the salient, semantically interesting features of the ML module system (and several proposed extensions to it) can be understood simply as stylized uses of a small set of orthogonal underlying constructs, with mixin composition playing a central role.We provide a declarative type system for MixML, including two important extensions: higher-order modules, and modules as first-class values. We also present a sound and complete, three-pass type-checking algorithm for this system. The operational semantics of MixML is defined by an elaboration translation into an internal core language called LTG---namely, a polymorphic lambda calculus with single-assignment references and recursive type generativity---which employs a linear type and kind system to track definedness of term and type imports.

Type checking and typability in domain-free lambda calculi

This paper shows (1) the undecidability of the type checking and the typability problems in the domain-free lambda calculus with negation, product, and existential types, (2) the undecidability of the typability problem in the domain-free polymorphic lambda calculus, and (3) the undecidability of the type checking and the typability problems in the domain-free lambda calculus with function and existential types. The first and the third results are proved by the second result and CPS translations that reduce those problems in the domain-free polymorphic lambda calculus to those in the domain-free lambda calculi with existential types. The key idea is the conservativity of the domain-free lambda calculi with existential types over the images of the translations.

Two extensions of system F with (co)iteration and primitive (co)recursion principles

This paper presents two extensions of the second order polymorphic lambda calculus, system F, with monotone (co)inductive types supporting (co)iteration, primitive (co)recursion and inversion principles as primitives. One extension is inspired by the usual categorical approach to programming by means of initial algebras and final coalgebras; whereas the other models dialgebras, and can be seen as an extension of Hagino's categorical lambda calculus within the framework of parametric polymorphism. The systems are presented in Curry-style, and are proven to be terminating and type-preserving. Moreover their expressiveness is shown by means of several programming examples, going from usual data types to lazy codata types such as streams or infinite trees.

Some Remarks on Type Systems for Course-of-value Recursion

Course-of-value recursion is a scheme which allows us to define the value of a function in some argument of an inductive structure by using not only the immediate, but arbitrary previously computed values. In the categorical approach to typed (total) functional programming, where datatypes (codatatypes) are initial algebras (final coalgebras), one models this principle by a construction called histomorphism. On the other hand, it is known that other categorical principles such as catamorphisms and anamorphisms representing definitions by iteration and coiteration have been successfully used to implement safe type systems extending the second-order polymorphic lambda calculus, system F. Hence it is natural to pursuit the definition and implementation of fold operators corresponding to course-of-value recursion as well. This paper proposes some new such extensions and states some important remarks emerged while verifying the correctness and safety properties of their operational semantics, relying not only on the categorical, but also on the logical approach based on fixed-point operators. Our observations should be considered as a starting point for a deeper study of the interrelation between these two approaches.

Typed self-representation

Self-representation -- the ability to represent programs in their own language -- has important applications in reflective languages and many other domains of programming language design. Although approaches to designing typed program representations for sublanguages of some base language have become quite popular recently, the question whether a fully metacircular typed self-representation is possible is still open. This paper makes a big step towards this aim by defining the F ω * calculus, an extension of the higher-order polymorphic lambda calculus F ω that allows typed self-representations. While the usability of these representations for metaprogramming is still limited, we believe that our approach makes a significant step towards a new generation of reflective languages that are both safe and efficient.

Selective strictness and parametricity in structural operational semantics, inequationally

Parametric polymorphism constrains the behavior of pure functional programs in a way that allows the derivation of interesting theorems about them solely from their types, i.e., virtually for free. The formal background of such ‘free theorems’ is well developed for extensions of the Girard–Reynolds polymorphic lambda calculus by algebraic datatypes and general recursion, provided the resulting calculus is endowed with either a purely strict or a purely nonstrict semantics. But modern functional languages like Clean and Haskell, while using nonstrict evaluation by default, also provide means to enforce strict evaluation of subcomputations at will. The resulting selective strictness gives the advanced programmer explicit control over evaluation order, but is not without semantic consequences: it breaks standard parametricity results. This paper develops an operational semantics for a core calculus supporting all the language features emphasized above. Its main achievement is the characterization of observational approximation with respect to this operational semantics via a carefully constructed logical relation. This establishes the formal basis for new parametricity results, as illustrated by several example applications, including the first complete correctness proof for short cut fusion in the presence of selective strictness. The focus on observational approximation, rather than equivalence, allows a finer-grained analysis of computational behavior in the presence of selective strictness than would be possible with observational equivalence alone.

The Girard–Reynolds isomorphism (second edition)

Jean-Yves Girard and John Reynolds independently discovered the second-order polymorphic lambda calculus, F2. Girard additionally proved a Representation Theorem: every function on natural numbers that can be proved total in second-order intuitionistic predicate logic, P2, can be represented in F2. Reynolds additionally proved an Abstraction Theorem: every term in F2 satisfies a suitable notion of logical relation; and formulated a notion of parametricity satisfied by well-behaved models. We observe that the essence of Girard’s result is a projection from P2 into F2, and that the essence of Reynolds’s result is an embedding of F2 into P2, and that the Reynolds embedding followed by the Girard projection is the identity. We show that the inductive naturals are exactly those values of type natural that satisfy Reynolds’s notion of parametricity, and as a consequence characterize situations in which the Girard projection followed by the Reynolds embedding is also the identity. An earlier version of this paper used a logic over untyped terms. This version uses a logic over typed term, similar to ones considered by Abadi and Plotkin and by Takeuti, which better clarifies the relationship between F2 and P2. This paper uses colour to enhance its presentation. If the link below is not blue, follow it for the colour version. http://homepages.inf.ed.ac.uk/wadler

A Type-theoretic Reconstruction of the Visitor Pattern

In object-oriented languages, the Visitor pattern can be used to traverse tree-like data structures: a visitor object contains some operations, and the data structure objects allow themselves to be traversed by accepting visitors. In the polymorphic lambda calculus (System F), tree-like data structures can be encoded as polymorphic higher-order functions. In this paper, we reconstruct the Visitor pattern from the polymorphic encoding by way of generics in Java. We sketch how the quantified types in the polymorphic encoding can guide reasoning about visitors in general.

Categorical models for Abadi and Plotkin's logic for parametricity

We propose a new category-theoretic formulation of relational parametricity based on a logic for reasoning about parametricity given by Abadi and Plotkin. The logic can be used to reason about parametric models, such that we may prove consequences of parametricity that to our knowledge have not been proved before for existing category-theoretic notions of relational parametricity. We provide examples of parametric models and describe a way of constructing parametric models from given models of the second-order lambda calculus.

Quantitative Information Flow, Relations and Polymorphic Types

This paper uses Shannon's information theory to give a quantitative definition of information flow in systems that transform inputs to outputs. For deterministic systems, the definition is shown to specialize to a simpler form when the information source and the known inputs jointly determine all inputs uniquely. For this special case, the definition is related to the classical security condition of non-interference and an equivalence is established between non-interference and independence of random variables. Quantitative information flow for deterministic systems is then presented in relational form. With this presentation, it is shown how relational parametricity can be used to derive upper and lower bounds on information flows through families of functions defined in the second-order lambda calculus.

An elementary fragment of second-order lambda calculus

A fragment of second-order lambda calculus (System F ) is defined that characterizes the elementary recursive functions. Type quantification is restricted to be noninterleaved and stratified, that is, the types are assigned levels, and a quantified variable can only be instantiated by a type of smaller level, with a slightly liberalized treatment of the level zero.

On proving the correctness of program transformations based on free theorems for higher-order polymorphic calculi

A number of program transformations currently of interest can be derived from Wadler's ‘free theorems’ for calculi approximating modern functional languages. Although delicate but fundamental issues arise in proving the correctness of free theorems-based program transformations, these issues have usually been left unaddressed in the correctness proofs appearing in the literature. As a result, most such proofs are incomplete, and most free theorems-based transformations are applied to programs in calculi for which they are not actually known to be correct.The purpose of this paper is three-fold. First, we raise and clarify some of the issues that must be addressed when constructing correctness proofs for free theorems-based program transformations. Second, we offer a principled approach to developing such proofs. Third, we use Pitts' recent work on parametricity and observational equivalence to show how our approach can be used to give the first proof that transformations based on the Acid Rain theorems preserve observational equivalence of programs in a polymorphic lambda calculus supporting FPC-style fixpoints and algebraic data types. Correctness of the foldr-build rule, the destroy-unfoldr rule, and the hylofusion program transformation for this calculus follows immediately. The same approach is expected to yield complete correctness proofs for free theorems-based transformations in calculi that even more closely resemble languages with which programmers are concerned in practice.