The three-continuation approach to coroutine pipelines efficiently represents a large number of connected components. Previous work in this area introduces this alternative encoding but does not shed much light on the underlying principles for deriving this encoding from its specification. This paper gives this missing insight by deriving the three-continuation encoding based on eliminating the mutual recursion in the definition of the connect operation. Using the same derivation steps, we are able to derive a similar encoding for a more general setting, namely bidirectional pipes. Additionally, we evaluate the encoding in an advertisement analytics benchmark where it is as performant as pipes, conduit, and streamly, which are other common Haskell stream processing libraries.

Algebraic effects and handlers are a convenient method for structuring monadic effects with primitive effectful operations and separating the syntax from the interpretation of these operations. However, the scope of conventional handlers is limited as not all side effects are monadic in nature. This paper generalizes the notion of algebraic effects and handlers from monads to generalized monoids, which notably covers applicative functors and arrows as well as monads. For this purpose, we switch the category theoretical basis from free algebras to free monoids. In addition, we show how lax monoidal functors enable the reuse of handlers and programs across different computation classes, for example, handling applicative computations with monadic handlers. We motivate and present these handler interfaces in the context of build systems. Tasks in a build system are represented by a free computation and their interpretation as a handler. This use case is based on the work of Mokhov et al. [(2018). PACMPL2(ICFP), 79:1–79:29.].

In ML-style module type theory, sealing often leads to situations in which type variables must leave scope, and this creates a need for signatures that avoid such variables. Unfortunately, in general, there is no best signature that avoids a variable, so modules do not always enjoy principal signatures. This observation is called the avoidance problem. In the past, the problem has been circumvented using a variety of devices for moving variables so they can remain in scope. These devices work, but have heretofore lacked a logical foundation. They have also lacked a presentation in which the dynamic semantics is given on the same phrases as the static semantics, which limits their applications. We can provide a best supersignature avoiding a variable by fiat, by adding an existential signature that is the least upper bound of its instances. This idea is old, but a workable metatheory has not previously been worked out. This work resolves the metatheoretic issues using ideas borrowed from focused logic. We show that the new theory results in a type discipline very similar to the aforementioned devices used in prior work. In passing, this gives a type-theoretic justification for the generative stamps used in the early days of the static semantics of ML modules. All the proofs are formalized in Coq.

In ‘Calculating Correct Compilers’ (Bahr & Hutton, 2015), we developed a new approach to calculating compilers directly from specifications of their correctness. Our approach only required elementary reasoning techniques and has been used to calculate compilers for a wide range of language features and their combination. However, the methodology was focused on stack-based target machines, whereas real compilers often target register-based machines. In this article, we show how our approach can naturally be adapted to calculate compilers for register machines.

Sheet-defined functions (SDFs) bring modularity and abstraction to the world of spreadsheets. Alas, end users naturally write SDFs that work over fixed-size arrays, which limits their reusability. To help end user programmers write more reusable SDFs, we describe a principled approach to generalising such functions to become elastic SDFs that work over inputs of arbitrary size. We prove that under natural, checkable conditions, our algorithm returns the principal generalisation of an input SDF. We describe a formal semantics and several efficient implementation strategies for elastic SDFs. A user study with spreadsheet users compares the human experience of programming with elastic SDFs to the alternative of relying on array-processing combinators. Our user study finds that the cognitive load of elastic SDFs is lower than for SDFs with map/reduce array combinators, the closest alternative solution.

Actors collaborate via message exchanges to reach a common goal. Experience has shown, however, that pure message-based communication is limiting and forces developers to use design patterns. The recently introduced dataspace actor model borrows ideas from the tuple space realm. It offers a tightly controlled, shared storage facility for groups of actors. In this model, actors assert facts that they wish to share and interests in such assertions. The dataspace notifies interested parties of changes to the set of assertions that they are interested in. Although it is straightforward to add the dataspace model to untyped languages, adding a typed interface is both necessary and challenging. Without restrictions on exchanged data, a faulty actor may propagate erroneous data through a malformed assertion, causing an otherwise well-behaved actor to crash—violating the key principle of failure isolation. A properly designed type system can prevent this scenario and rule out other kinds of uncooperative actors. This paper presents the first structural type system for the dataspace model of actors; it does not address the question of behavioral types for assertion-oriented protocols.

Many types of interactive applications, including reactive systems implemented in hardware, interactive physics simulations and games, raise particular challenges when it comes to testing and debugging. Reasons include de facto lack of reproducibility and difficulties of automatically generating suitable test data. This paper demonstrates that certain variants of functional reactive programming (FRP) implemented in pure functional languages can mitigate such difficulties by offering referential transparency at the level of whole programs. This opens up for a multi-pronged approach for assisting with testing and debugging that works across platforms, including assertions based on temporal logic, recording and replaying of runs (also from deployed code), and automated random testing using QuickCheck. When combined with extensible forms of FRP that allow for constrained side effects, it allows us to not only validate software simulations but to analyse the effect of faults in reactive systems, confirm the efficacy of fault tolerance mechanisms and perform software- and hardware-in-the-loop testing. The approach has been validated on non-trivial systems implemented in several existing FRP implementations, by means of careful debugging using a tool that allows the test or simulation under scrutiny to be controlled, moving along the execution time line, and pin-pointing of violations of assertions on personal computers as well as external devices.