Context-oriented programming treats execution context explicitly and provides means for context-dependent adaptation at runtime. This is achieved, for example, using dynamic layer activation and contextual dispatch, where the context consists of a layer environment of a stack of active layers. Layers can adapt existing behaviour using proceed to access earlier activated layers. A problem arises when a call to proceed is made from within a closure that escapes the layer environment in which it was defined. It is not clear how to interpret proceed when the closure is subsequently applied in a different environment, because the layers it implicitly refers to (such as the original layer and/or the remaining layers) may no longer be active. In this paper, we describe this problem in detail and present some approaches for dealing with it, though ultimately we leave the question open. 1.

It is often hard to write programs that are efficient yet reusable. For example, an efficient implementation of Gaussian elimination should be specialized to the structure and known static properties of the input matrix. The most profitable optimizations, such as choosing the best pivoting or memoization, cannot be expected of even an advanced compiler because they are specific to the domain, but expressing these optimizations directly makes for ungainly source code. Instead, a promising and popular way to reconcile efficiency with reusability is for a domain expert to write code generators. Two pillars of this approach are types and effects. Typed multilevel languages such as MetaOCaml ensure safety: a well-typed code generator neither goes wrong nor generates code that goes wrong. Side effects such as state and control ease correctness: an effectful generator can resemble the textbook presentation of an algorithm, as is familiar to domain experts, yet insert let for memoization and if for bounds-checking, as is necessary for efficiency. However, adding effects blindly renders multilevel types unsound. We introduce the first two-level calculus with control effects and a sound type system. We give small-step operational semantics as well as a continuation-passing style (CPS) translation. For soundness, our calculus restricts the code generator’s effects to the scope of generated binders. Even with this restriction, we can finally write efficient code generators for dynamic programming and numerical methods in direct style, like in algorithm textbooks, rather than in CPS or monadic style.

We define a type system with intersection types for an extension of lambda-calculus with unbind and rebind operators. In this calculus, a term t with free variables x1,...,xn, representing open code, can be packed into an unbound term 〈 x1,...,xn | t 〉, and passed around as a value. In order to execute inside code, an unbound term should be explicitly rebound at the point where it is used. Unbinding and rebinding are hierarchical, that is, the term t can contain arbitrarily nested unbound terms, whose inside code can only be executed after a sequence of rebinds has been applied. Correspondingly, types are decorated with levels, and a term has type τ k if it needs k rebinds in order to reduce to a value of type τ. With intersection types we model the fact that a term can be used differently in contexts providing a different numbers of unbinds. In particular, top-level terms, that is, terms not requiring unbinds to reduce to values, should have a value type, that is, an intersection type where at least one element has level 0. With the proposed intersection type system we get soundness under the call-by-value strategy, an issue which was not resolved by previous type systems.

We describe the implementation of first-class polymorphic delimited continuations in the programming language Scala. We use Scala’s pluggable typing architecture to implement a simple type and effect system, which discriminates expressions with control effects from those without and accurately tracks answer type modification incurred by control effects. To tackle the problem of implementing first-class continuations under the adverse conditions brought upon by the Java VM, we employ a selective CPS transform, which is driven entirely by effect-annotated types and leaves pure code in direct style. Benchmarks indicate that this high-level approach performs competitively.

Abstract. We extend the simply typed λ-calculus with unbind and rebind primitive constructs. That is, a value can be a fragment of open code, which in order to be used should be explicitly rebound. This mechanism nicely coexists with standard static binding. The motivation is to provide an unifying foundation for mechanisms of dynamic scoping, where the meaning of a name is determined at runtime, rebinding, such as dynamic updating of resources and exchange of mobile code, and delegation, where an alternative action is taken if a binding is missing. Depending on the application scenario, we consider two extensions which differ in the way type safety is guaranteed. The former relies on a combination of static and dynamic type checking. That is, rebind raises a dynamic error if for some variable there is no replacing term or it has the wrong type. In the latter, this error is prevented by a purely static type system, at the price of more sophisticated types. 1991 Mathematics Subject Classification. 68N15, 68N18.

Abstract. Concurrency is fundamental to both the expressivity and efficiency of programs today. Systems of multiple concurrent processes require coordination. In the actor model, these coordination constructs are often designed in terms of meta-actors. But there are actually many different approaches to designing an actor meta-architecture, which in turn affect how coordination constructs can be designed. In this paper, we consider the application of request-based reflection to prototyping actor meta-architectures. We define the actor model in terms of a reflective language, and consider how this technique differs from other approaches. We then use the reflective language to introduce additional layers of functionality: local synchronization constraints, statically-assigned meta-actors, and dynamically-reassignable meta-actors. The examples indicate that request-based reflection is useful for prototyping a variety of different meta-architectural constructs. 1

Custom code generation is the leading approach to reconciling generality with performance. MetaOCaml, a dialect of OCaml, is the best-developed way today to write custom code generators and assure them type-safe across multiple stages of computation. Nevertheless, the continuing interest from the community has yet to result in a mature implementation of MetaOCaml that integrates cleanly with OCaml’s type checker and run-time system. Even in theory, it is unclear how staging interacts with effects, polymorphism, and user-defined data types. We report on the status of the ongoing MetaOCaml project, focusing on the gap between theory and practice and the difficulties that arise in a full-featured staged language rather than an idealized calculus. We highlight foundational problems in type soundness and cross-stage persistence that demand investigation. We also suggest a lightweight implementation of a two-stage dialect of OCaml, as syntactic sugar. 1.

We describe the first implementation of multi-prompt delimited control operators in OCaml that is direct in that it captures only the needed part of the control stack. The implementation is a library that requires no changes to the OCaml compiler or run-time, so it is perfectly compatible with existing OCaml source and binary code. The library has been in fruitful practical use since 2006. We present the library as an implementation of an abstract machine derived by elaborating the definitional machine. The abstract view lets us distill a minimalistic API, scAPI, sufficient for implementing multi-prompt delimited control. We argue that a language system that supports exception and stack-overflow handling supports scAPI. With byte- and native-code OCaml systems as two examples, our library illustrates how to use scAPI to implement multi-prompt delimited control in a typed language. The approach is general and has been used to add multi-prompt delimited control to other existing language systems.