Claudio V. Russo

Types for Modules

Abstract The programming language Standard ML is an amalgam of two, largely orthogonal, languages. The Core language expresses details of algorithms and data structures. The Modules language expresses the modular architecture of a software system. Both languages are statically typed, with their static and dynamic semantics specified by a formal definition. Over the past decade, Standard ML Modules has been the source of inspiration for much research into the type-theoretic foundations of modules languages. Despite these efforts, a proper type-theoretic understanding of its static semantics has remained elusive. In this thesis, we use Type Theory as a guideline to reformulate the unconventional static semantics of Modules, providing a basis for useful extensions to the Modules language. Our starting point is a stylised presentation of the existing static semantics of Modules, parameterised by an arbitrary Core language. We claim that the type-theoretic concepts underlying Modules are type parameterisation, type quantification and subtyping. We substantiate this claim by giving a provably equivalent semantics with an alternative, more type-theoretic presentation. In particular, we show that the notion of type generativity corresponds to existential quantification over types. In contrast to previous accounts, our analysis does not involve first-order dependent types. Our first extension generalises Modules to higher-order, allowing modules to take parameterised modules as arguments, and return them as results. We go beyond previous proposals for higher-order Modules by supporting a notion of type generativity. We give a sound and complete algorithm for type-checking higher-order Modules. Our second extension permits modules to be treated as first-class citizens of an ML-like Core language, greatly extending the range of computations on modules. Each extension arises from a natural generalisation of our type-theoretic semantics. This thesis also addresses two pragmatic concerns. First, we propose a simple approach to the separate compilation of Modules, which is adequate in practice but has theoretical limitations. We suggest a modified syntax and semantics that alleviates these limitations. Second, we study the type inference problem posed by uniting our extensions to higher-order and first-class modules with an implicitly-typed, ML-like Core language. We present a hybrid type inference algorithm that integrates the classical algorithm for ML with the type-checking algorithm for Modules.

Operational Properties of Lily, a Polymorphic Linear Lambda Calculus with Recursion

Abstract Plotkin has advocated the combination of linear lambda calculus, polymorphism and fixed point recursion as an expressive semantic metalanguage. We study its expressive power from an operational point of view. We show that the naturally call-by-value operators of linear lambda calculus can be given a call-by-name semantics without affecting termination at exponential types and hence without affecting ground contextual equivalence. This result is used to prove properties of a logical relation that provides a new extensional characterisation of ground contextual equivalence and relational parametricity properties of polymorphic types.

Variance and generalized constraints for C # generics

Generic types in C

Generalized algebraic data types and object-oriented programming

^{\sharp}

Tabular: a schema-driven probabilistic programming language

behave invariantly with respect to subtyping. We propose a system of type-safe variance for C

Parallel concurrent ML

^{\sharp}

Adventures in interoperability: the SML.NET experience

that supports the declaration of covariant and contravariant type parameters on generic types. To support more widespread application of variance we also generalize the existing constraint mechanism with arbitrary subtype assertions on classes and methods. This extension is useful even in the absence of variance, and subsumes equational constraints proposed for Generalized Algebraic Data Types (GADTs). We formalize the subtype relation in both declarative and syntax-directed style, and describe and prove the correctness of algorithms for constraint closure and subtyping. Finally, we formalize and prove a type safety theorem for a featherweight language with variant classes and generalized constraints.

F-ing Modules

Generalized algebraic data types (GADTs) have received much attention recently in the functional programming community. They generalize the (type) parameterized algebraic datatypes (PADTs) of ML and Haskell by permitting value constructors to return specific, rather than parametric, type-instantiations of their own datatype. GADTs have a number of applications, including strongly-typed evaluators, generic pretty-printing, generic traversals and queries, and typed LR parsing. We show that existing object-oriented programming languages such as Java and C# can express GADT definitions, and a large class of GADT-manipulating programs, through the use of generics, subclassing, and virtual dispatch. However, some programs can be written only through the use of redundant runtime casts. Moreover, instantiation-specific, yet safe, operations on ordinary PADTs only admit indirect cast-free implementations, via higher-order encodings. We propose a generalization of the type constraint mechanisms of C# and Java to both avoid the need for casts in GADT programs and higher-order contortions in PADT programs; we present a Visitor pattern for GADTs, and describe a refined switch construct as an alternative to virtual dispatch on datatypes. We formalize both extensions and prove type soundness.

Deriving probability density functions from probabilistic functional programs

We propose a new kind of probabilistic programming language for machine learning. We write programs simply by annotating existing relational schemas with probabilistic model expressions. We describe a detailed design of our language, Tabular, complete with formal semantics and type system. A rich series of examples illustrates the expressiveness of Tabular. We report an implementation, and show evidence of the succinctness of our notation relative to current best practice. Finally, we describe and verify a transformation of Tabular schemas so as to predict missing values in a concrete database. The ability to query for missing values provides a uniform interface to a wide variety of tasks, including classification, clustering, recommendation, and ranking.

Pause 'n' play: formalizing asynchronous C#

Concurrent ML (CML) is a high-level message-passing language that supports the construction of first-class synchronous abstractions called events. This mechanism has proven quite effective over the years and has been incorporated in a number of other languages. While CML provides a concurrent programming model, its implementation has always been limited to uniprocessors. This limitation is exploited in the implementation of the synchronization protocol that underlies the event mechanism, but with the advent of cheap parallel processing on the desktop (and laptop), it is time for Parallel CML. Parallel implementations of CML-like primitives for Java and Haskell exist, but build on high-level synchronization constructs that are unlikely to perform well. This paper presents a novel, parallel implementation of CML that exploits a purpose-built optimistic concurrency protocol designed for both correctness and performance on shared-memory multiprocessors. This work extends and completes an earlier protocol that supported just a strict subset of CML with synchronization on input, but not output events. Our main contributions are a model-checked reference implementation of the protocol and two concrete implementations. This paper focuses on Manticores functional, continuation-based implementation but briefly discusses an independent, thread-based implementation written in C# and running on Microsofts stock, parallel runtime. Although very different in detail, both derive from the same design. Experimental evaluation of the Manticore implementation reveals good performance, dispite the extra overhead of multiprocessor synchronization.