Compiler Theory Research Articles

Instruction and Practice in Learning to use a Device

We explore the extent to which Anderson's (1987) theory of knowledge compilation can account for the relationship between instructions and practice in learning to use a simple device. Bibby and Payne (1993) reported experimental support for knowledge compilation in this domain. This article replicates the finding of a performance cross‐over between instruction type and task type that disappears with practice on the tasks. The research is extended by using verbal protocols to model the strategies of novice and more experienced individuals. Production system models of these strategies suggest that knowledge compilation only provides an adequate account of practice for one of two Instruction groups. To model the strategy shifts for the second group, we employ a “procedure modification heuristic” (after Neches, 1987), which relies on access to a declarative model of the configuration of the device. This suggests that instructionally derived declarative knowledge may have a subtle ongoing effect on the changes in procedural knowledge with practice.

Enriching the lambda calculus with contexts

A context in the λ-calculus is a term with some holes. Hole filling differs from β-substitution in that name capture is intended. This seemingly simple feature transcends static scope and lies at the heart of modular and object-oriented programming. Still, the name capture feature of hole filling is at odds with hygienic β-substitution. In this paper we conservatively extend the λ-calculus to incorporate the notion of contexts without jeopardizing the β-rule. We perceive contexts as source code and λ-terms as target code. Context filling is encoded as compilation operations and the enriched calculus is a theory of separate compilation and incremental program construction. Linking of separately-developed programs is done by coherent renaming of free variables.We apply our context-enriching schema to the λ-calculus extended with definitions and devise a calculus of first-class modules. We show that module linking can be modeled solely by the renaming of import and export variables. We add relinkable variable references to model virtual method references essential to object systems.The inclusion of contexts introduces parameters whose linking is based on names (symbols, identifiers, or keywords). We simulate in the context-enriched calculus other extensions of the λ-calculus with name-based programming notions such as Dami's λ-calculus with names, Aït-Kaci and Garrigue's label-selective λ-calculus, Lamping's transparent data parameters, and our quasi-static procedures.

Automated dismemberment of JSD process specifications

In the initial phase of Jackson System Development (JSD), behaviour in the real world is modelled using long-running processes. As direct implementation of such processes can be impractical, a transformational approach called dismemberment is sometimes applied to allow the separate scheduling of portions of a process's text. A lack of automatic support has, however, made the technique difficult to use, and automation itself has not proved easy. Several writers have noted the relationship between Jackson structure diagrams and finite automata. This paper describes in detail the transformation of JSD process specifications into deterministic finite automata, using ‘subset construction’, a method adapted from compiler theory. The resulting representations can be used to generate dismembered implementation routines. The algorithms have been implemented within an experimental version of the PRESTIGE JSD Workbench, a CASE tool for the JSD implementor.

Production probability estimators for context-free grammars

The problem of production probability estimation is considered. We examine the estimation of production probabilities for context-free probabilistic grammars (CFPGs). Given a context-free grammar, G, and a random sample of strings from L( G), we define ratio estimators to estimate the production probabilities of G. It is shown that the statistical analysis of the estimators becomes much less complex by using the theory of random walks. The main result of the article is that the biases and variances of all ratio estimators for any Chomsky Normal Form CFPG are approximately directly proportional to 1 n , where n is the size of the random sample. Random samples consisting of strings generated from an independent source are used to validate theoretical results. We show that the estimates can be used to increase the efficiency of parsing strings in a language. The ability to parse strings efficiently is extremely useful in compiler theory and the theory of artificial intelligence.

Theory and practice of intensional compilation of queries: Michel de Rougemont, European computer-industry research center, Muenchen, FRG, 1985

Theory and practice of intensional compilation of queries: Michel de Rougemont, European computer-industry research center, Muenchen, FRG, 1985

Toward compiler implementation correctness proofs

Aspect of the interaction between compiler theory and practice is addressed. Presented is a technique for the syntax-directed specification of compilers together with a method for proving the correctness of their parse-driven implementations. The subject matter is presented in an order-algebraic framework; while not strictly necessary, this approach imposes beneficial structure and modularity on the resulting specifications and implementation correctness proofs. Compilers are specified using an order-algebraic definition of attribute grammars. A practical class of compiler implementations is considered, consisting of those driven by LR( k ) or LL( k ) parsers which cause a sequence of translation routine activations to modify a suitably initialized collection of data structures (called a translation environment). The implementation correctness criterion consists of appropriately comparing, for each source program, the corresponding object program (contained in the final translation environment) produced by the compiler implementation to the object program dictated by the compiler specification. Provided that suitable intermediate assertions (called translation invariants ) are supplied, the program consisting of the (parse-induced) sequence of translation routine activations can be proven partially correct via standard inductive assertion methods.

E. K. Blum. Towards a theory of semantics and compilers for programming languages. Journal of computer and system sciences, vol. 3 (1969), pp. 248–275.

E. K. Blum. Towards a theory of semantics and compilers for programming languages. Journal of computer and system sciences, vol. 3 (1969), pp. 248–275.

Industry reaction to computer science education

Today, undergraduate Computer Science Education teaches mechanics without teaching problem solving. Typical curricula include courses in assembler languages, compiler theory, list processing, and automata theory. Every MS degree holder, and most BS degree holders, know Polish notation and have written parts of compilers. However, few of them have ever learned to write a program that can be easily enhanced or respond to changes as new management (instructor) requirements are set forth. Even fewer can read a program and describe what it does or debug a system consisting of ten or more modules.

Towards a theory of semantics and compilers for programming languages

The concept of imbedding a programming language L into a formal system is introduced and used as the basis of defining the semantics Θ of the language. Θ is an operator which maps a program p in L onto a function. The pair (L, Θ) is called a programming system. If (L, Θ) and (L,' Θ') are two programming systems, a mapping Γ from L into L' is called a compiler. A compiler is said to be correct if Θ'(Γ(p)) = Θ(p) for p in L. These ideas are illustrated in terms of two programming systems, one for recursive functions and the other for Turing machines.