Bernd Braßel

KiCS2: a new compiler from Curry to Haskell

In this paper we present our first steps towards a new system to compile functional logic programs of the source language Curry into purely functional Haskell programs. Our implementation is based on the idea to represent non-deterministic results as values of the data types corresponding to the results. This enables the application of various search strategies to extract values from the search space. We show by several benchmarks that our implementation can compete with or outperform other existing implementations of Curry.

On a tighter integration of functional and logic programming

The integration of functional and logic programming is a well developed field of research. We discuss that the integration could be improved significantly in two separate aspects: sharing computations across non-deterministic branching and the declarative encapsulation of logic search. We then present a formal approach which shows how these improvements can be realized and prove the soundness of our approach.

The Kiel Curry System KiCS

This paper presents the Kiel Curry System ( KiCS ), a new implementation of the lazy functional logic language Curry. Its main features beyond other Curry implementations are: flexible search control by means of search trees, referentially transparent encapsulation and sharing across non-determinism.

Observing Functional Logic Computations

A lightweight approach to debugging functional logic programs by observations is presented, implemented for the language Curry. The Curry Object Observation System (COOSy) comprises a portable library plus a viewing tool. A programmer can observe data structures and functions by annotating expressions in his program. The possibly partial values of observed expressions that are computed during program execution are recorded in a trace file, including information on non-deterministic choices and logical variables. A separate viewing tool displays the trace content. COOSy covers all aspects of modern functional logic multiparadigm languages such as lazy evaluation, higher order functions, non-deterministic search, logical variables, concurrency and constraints. Both use and implementation of COOSy are described.

High-level database programming in curry

This paper presents an environment to support high-level database programming in the multi-paradigm declarative programming language Curry. We define an application programming interface (API) that abstracts from the concrete database access methods. The API supports transactions and exploits Currys type system to ensure a strict separation between queries and updates. In order to ensure database updates that are safe w.r.t. an intended data model (e.g., containing specific relations between entities), we assume a description of the data dependencies in the entity-relationship (ER) model from which all access and update operations related to the database are generated. We propose a representation of ER diagrams in the declarative language Curry so that they can be constructed by various tools and then translated into this representation. Furthermore, we have implemented a compiler from this representation into a Curry program that provides safe access and update operations based on the API for database programming.

Transforming functional logic programs into monadic functional programs

We present a high-level transformation scheme to translate lazy functional logic programs into pure Haskell programs. This transformation is based on a recent proposal to efficiently implement lazy non-deterministic computations in Haskell in a monadic style. We build on this work and define a systematic method to transform lazy functional logic programs into monadic programs with explicit sharing. This results in a transformation scheme which produces high-level and flexible target code. For instance, the target code is parametric w.r.t. the concrete evaluation monad. Thus, different monad instances could, for example, define different search strategies (e.g., depth-first, breadth-first, parallel). We formally describe the basic compilation scheme and some useful extensions.

Declaring Numbers

Most implementations of functional and functional logic languages treat numbers and the basic numeric operations as external entities. The main reason for this is efficiency. However, this basic design decision has many unfortunate consequences for all programs using numbers. We present an approach to model numbers in a declarative way and argue that the loss in efficiency is compensated by the newly gained possibilities. Functional logic languages benefit the most from this proposal because all the numeric operations become fully narrowable. This enables the solving of simple equations on numbers in an efficient way without having to resort to external constraint solvers. The presented approach can either be used as a library for purely declarative numbers or it can be employed as a basic data type of functional (logic) languages. Indeed, we have integrated the presented data structures as the only numbers available in our compiler for the functional logic language Curry.

Lazy call-by-value evaluation

Designing debugging tools for lazy functional programming languages is a complex task which is often solved by expensive tracing of lazy computations. We present a new approach in which the information collected as a trace is reduced considerably (kilobytes instead of megabytes). The idea is to collect a kind of step information for a call-by-value interpreter, which can then efficiently reconstruct the computation for debugging/viewing tools, like declarative debugging. We show the correctness of the approach, discuss a proof-of-concept implementation with a declarative debugger as back end and present some benchmarks comparing our new approach with the Haskell debugger Hat.

Computing with subspaces

We propose a new definition and use of a primitive getAllValues, for computing all the values of a non-deterministic expression in a functional logic program. Our proposal restricts the validity of the argument of getAllValues. This restriction ensures that essential language features like the call-time choice semantics, the independence of the order of evaluation, and the referential transparency of the language are preserved when getAllValues is executed. Up to now, conflicts between these language features and primitives like getAllValues have been seen as one of the main problems for employing such primitives in functional logic languages.

Nondeterminism analysis of functional logic programs

Information about the nondeterminism behavior of a functional logic program is important for various reasons. For instance, a nondeterministic choice in I/O operations results in a run-time error. Thus, it is desirable to ensure at compile time that a given program is not going to crash in this way. Furthermore, knowledge about nondeterminism can be exploited to optimize programs. In particular, if functional logic programs are compiled to target languages without builtin support for nondeterministic computations, the transformation can be much simpler if it is known that the source program is deterministic. In this paper we present a nondeterminism analysis of functional logic programs in form of a type/effect system. We present a type inferencer to approximate the nondeterminism behavior via nonstandard types and show its correctness w.r.t. the operational semantics of functional logic programs. The type inference is based on a new compact representation of sets of types and effects.