Nils Anders Danielsson

Fast and loose reasoning is morally correct

Functional programmers often reason about programs as if they were written in a total language, expecting the results to carry over to non-total (partial) languages. We justify such reasoning.Two languages are defined, one total and one partial, with identical syntax. The semantics of the partial language includes partial and infinite values, and all types are lifted, including the function spaces. A partial equivalence relation (PER) is then defined, the domain of which is the total subset of the partial language. For types not containing function spaces the PER relates equal values, and functions are related if they map related values to related values.It is proved that if two closed terms have the same semantics in the total language, then they have related semantics in the partial language. It is also shown that the PER gives rise to a bicartesian closed category which can be used to reason about values in the domain of the relation.

Lightweight semiformal time complexity analysis for purely functional data structures

Okasaki and others have demonstrated how purely functional data structures that are efficient even in the presence of persistence can be constructed. To achieve good time bounds essential use is often made of laziness. The associated complexity analysis is frequently subtle, requiring careful attention to detail, and hence formalising it is valuable. This paper describes a simple library which can be used to make the analysis of a class of purely functional data structures and algorithms almost fully formal. The basic idea is to use the type system to annotate every function with the time required to compute its result. An annotated monad is used to combine time complexity annotations. The library has been used to analyse some existing data structures, for instance the deque operations of Hinze and Patersons finger trees.

Total parser combinators

A monadic parser combinator library which guarantees termination of parsing, while still allowing many forms of left recursion, is described. The librarys interface is similar to those of many other parser combinator libraries, with two important differences: one is that the interface clearly specifies which parts of the constructed parsers may be infinite, and which parts have to be finite, using dependent types and a combination of induction and coinduction; and the other is that the parser type is unusually informative. The library comes with a formal semantics, using which it is proved that the parser combinators are as expressive as possible. The implementation is supported by a machine-checked correctness proof.

Operational semantics using the partiality monad

The operational semantics of a partial, functional language is often given as a relation rather than as a function. The latter approach is arguably more natural: if the language is functional, why not take advantage of this when defining the semantics? One can immediately see that a functional semantics is deterministic and, in a constructive setting, computable. This paper shows how one can use the coinductive partiality monad to define big-step or small-step operational semantics for lambda-calculi and virtual machines as total, computable functions (total definitional interpreters). To demonstrate that the resulting semantics are useful type soundness and compiler correctness results are also proved. The results have been implemented and checked using Agda, a dependently typed programming language and proof assistant.

A formalisation of a dependently typed language as an inductive-recursive family

It is demonstrated how a dependently typed lambda calculus (a logical framework) can be formalised inside a language with inductiverecursive families. The formalisation does not use raw terms; the welltyped terms are defined directly. It is hence impossible to create ill-typed terms. As an example of programming with strong invariants, and to show that the formalisation is usable, normalisation is proved. Moreover, this proof seems to be the first formal account of normalisation by evaluation for a dependently typed language.

Beating the Productivity Checker Using Embedded Languages

Some total languages, like Agda and Coq, allow the use of guarded corecursion to construct infinite values and proofs. Guarded corecursion is a form of recursion in which arbitrary recursive calls are allowed, as long as they are guarded by a coinductive constructor. Guardedness ensures that programs are productive, i.e. that every finite prefix of an infinite value can be computed in finite time. However, many productive programs are not guarded, and it can be nontrivial to put them in guarded form. This paper gives a method for turning a productive program into a guarded program. The method amounts to defining a problem-specific language as a data type, writing the program in the problemspecific language, and writing a guarded interpreter for this language.

ΠΣ: dependent types without the sugar

The recent success of languages like Agda and Coq demonstrates the potential of using dependent types for programming. These systems rely on many high-level features like datatype definitions, pattern matching and implicit arguments to facilitate the use of the languages. However, these features complicate the metatheoretical study and are a potential source of bugs. To address these issues we introduce ΠΣ, a dependently typed core language. It is small enough for metatheoretical study and the type checker is small enough to be formally verified. In this language there is only one mechanism for recursion—used for types, functions and infinite objects—and an explicit mechanism to control unfolding, based on lifted types. Furthermore structural equality is used consistently for values and types; this is achieved by a new notion of α-equality for recursive definitions. We show, by translating several high-level constructions, that ΠΣ is suitable as a core language for dependently typed programming.

Parsing mixfix operators

A simple grammar scheme for expressions containing mixfix operators is presented. The scheme is parameterised by a precedence relation which is only restricted to be a directed acyclic graph; this makes it possible to build up precedence relations in a modular way. Efficient and simple implementations of parsers for languages with user-defined mixfix operators, based on the grammar scheme, are also discussed. In the future we plan to replace the support for mixfix operators in the language Agda with a grammar scheme and an implementation based on this work.

Matroids from Modules

Abstract The aim of this work is to show that (oriented) matroid methods can be applied to many discrete geometries, namely those based on modules over integral (ordered) domains. The trick is to emulate the structure of a vector space within the module, thereby allowing matroid methods to be used as if the module were a vector space. Only those submodules which are “closed under existing divisors”, and hence behave like vector subspaces, are used as subspaces of the matroid. It is also shown that Hubler‘s axiomatic discrete geometry can be characterised in terms of modules over the ring of integers.

Bag Equivalence via a Proof-Relevant Membership Relation

Two lists are bag equivalent if they are permutations of each other, i.e. if they contain the same elements, with the same multiplicity, but perhaps not in the same order. This paper describes how one can define bag equivalence as the presence of bijections between sets of membership proofs. This definition has some desirable properties: Many bag equivalences can be proved using a flexible form of equational reasoning. The definition generalises easily to arbitrary unary containers, including types with infinite values, such as streams. By using a slight variation of the definition one gets set equivalence instead, i.e. equality up to order and multiplicity. Other variations give the subset and subbag preorders. The definition works well in mechanised proofs.