Declarative Programming Sasbachwalden 1991

Declarative programming languages are based on sound mathematical foundations which means that they offer many advantages for software development. These advantages include their powerful descriptive capabilities, the availability of program analysis techniques and the potential for parallel execution. This volume contains the proceedings of a seminar and workshop organised by the Esprit Basic Research Action Phoenix in collaboration with the Esprit Basic Research Action Integration. Both these groups have been closely involved in investigating the foundations of declarative programming and the integration of various language paradigms, as well as the developing aspects of related technology. The main aim of the seminar and workshop was to provide a forum for the results of this work, together with contributions from other researchers in the same field. These papers cover a variety of important technical areas such as foundations and languages, program transformation and analysis, integrated approaches, implementation techniques, abstract machines and programming methodology. The resulting volume provides an in-depth picture of current research into declarative programming. It will be of special interest to researchers in programming languages and methodology, students of artificial intelligence and anyone involved in industrial research and development.

The Glasgow functional programming group has held a workshop each summer since 1988. The entire group, accompanied by a selection of colleagues from other institutions, retreats to a pleasant Scottish location for a few days. Everyone speaks briefly, enhancing coherence, cross fertilisation, and camaraderie in our work. The proceedings of the first workshop were published as a technical report. Demand for this was large enough to encourage wider publication, and subsequent proceedings have been published in the Springer-Verlag Workshops in Computing series. These are the proceedings of the-meeting held 12-14 August 1991, in Portree on the Isle of Skye. A preliminary proceedings was prepared in advance of the meeting. Most presentations were limited to a brief fifteen minutes, outlining the essentials of their subject, and referring the audience to the pre-print proceedings for details. Papers were then refereed and rewritten, and you hold the final results in your hands. A number of themes emerged at this year's workshop, including relational algebra and its application to hardware design, partial evaluation and program transformation, implementation techniques, and strictness analysis. We were especially pleased to see applications of functional programming emerge as a theme. One of the sessions was devoted to a lively discussion of applications, and was greatly enhanced by our industrial participants. The workshop was organised by Kei Davis, Cordelia Hall, Rogardt Heldal, Carsten Kehler Holst, John Hughes, John O'Donnell, and Satnam Singh all from the University of Glasgow.

In ordinary mathematics, an equation can be written down which is syntactically correct, but for which no solution exists. For example, consider the equation x = x + 1 defined over the real numbers; there is no value of x which satisfies it. Similarly it is possible to specify objects using the formal specification language Z [3,4], which can not possibly exist. Such specifications are called inconsistent and can arise in a number of ways. Example 1 The following Z specification of a functionf, from integers to integers "f x : ~ 1 x ~ O· fx = x + 1 (i) "f x : ~ 1 x ~ O· fx = x + 2 (ii) is inconsistent, because axiom (i) gives f 0 = 1, while axiom (ii) gives f 0 = 2. This contradicts the fact that f was declared as a function, that is, f must have a unique result when applied to an argument. Hence no suchfexists. Furthermore, iff 0 = 1 andfO = 2 then 1 = 2 can be deduced! From 1 = 2 anything can be deduced, thus showing the danger of an inconsistent specification. Note that all examples and proofs start with the word Example or Proof and end with the symbol.1.

This book contains the edited versions of papers presented at the Fourth Irish Conference on Artificial Intelligence and Cognitive Science (AICS'91), which was held at University College, Cork, Ireland on 19-20 September 1991. The main aims of this annual conference series are to promote AI research in Ireland, to provide a forum for the exchange of ideas amongst different disciplines concerned with the study of cognition, and to provide an opportunity for industry to see what research is being carried out in Ireland and how it might benefit from the results of this research. While most of the participants at the conference came from universities and companies within Ireland, a positive feature was the interest shown from outside the country, resulting in participants from Britain, USA and Italy. The keynote speaker was Professor James A. Bowen, North Carolina State University, who spoke on future trends in knowledge representation. The topics covered in the presented papers included fundamental approaches to AI, natural language, knowledge representation, information retrieval, deduction, epistemics and vision. The sponsors of the conference were Digital Equipment Co. (Galway) and Eolas, the Irish Science and Technology Board. March 1992 Humphrey Sorensen Contents Section 1: Foundations and Methodologies Toward a New Foundation for Cognitive Science S. 6 Nuallain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Towards an Adequate Cognitive Model of Analogical Mapping M. T. Keane and S. Duff. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 HOST: A HOlistic System Theory G. Hartnett and H. Sorensen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Glasgow Functional Programming Group is widely recognised for its research in lazy functional languages. Once again this year, for the fifth time, we retreated to a Scottish seaside town to discuss our latest work, this time spending three days in Ayr. We were joined by a number of colleagues from other universities and from industry, with whom we have been enjoying fruitful collaboration. The workshop serves the dual purpose of ensuring that the whole group remains informed of each other's work, and of providing workshop experience for research students. Most participants presented a short talk about their work, supplemented by papers which appeared in a draft proceedings distributed at the workshop. Since then the papers have been reviewed and the majority are now published here following revision. The workshop also contained a lively discussion session on functional language applications, to which the industrial participants made very helpful contributions. One interesting feature of this volume is the number of papers addressing practical issues of realistic use of functional languages, from benchmarking and profiling, to user interfaces and file handling. It is perhaps indicative that at last lazy functional languages are being used for significantly larger applications than has been typical in the past. This type of paper is likely to feature prominently in relevant conference proceedings for the next few years. Other papers here continue to address more theoretical topics, such as program logic, semantics of non determinism, program analysis, and loop detection.

Logic program synthesis and transformation are topics of central importance to the software industry. The demand for software can not be met by the current supply, in terms of volume, complexity, or reliability. The most promising solution seems to be the increased automation of software production: programmer productivity would improve, and correctness could be ensured by the application of mathematical methods. Because of their mathematical foundations, logic programs lend themselves particularly well to machine-assisted development techniques, and therefore to automation. This volume contains the proceedings of the second International Workshop on Logic Program Synthesis and Transformation (LOPSTR 92), held at the University of Manchester, 2-3 July 1992. The LOPSTR workshops are the only international meetings devoted to these two important areas. A variety of new techniques were described at the workshop, all of which promise to revolutionize the software industry once they become standard practise. These include techniques for the transformation of an inefficient program into an equivalent, efficient one, and the synthesis of a program from a formal specification of its required behaviour. Among the topics covered in this volume are: optimal transformation of logic programs; logic program synthesis via proof planning; deductive synthesis of programs for query answering; efficient compilation of lazy narrowing into Prolog; synthesis of narrowing programs; Logimix: a self-applicable partial evaluator for Prolog; proof nets; automatic termination analysis. Logic Program Synthesis and Transformation describes the latest advances in machine-assisted development of logic programs. It will provide essential reading for researchers and postgraduate students concerned with these two important areas.

This is the latest volume in the 'Workshops in Computing' series, and contains papers from the International Workshop on Hpyermedia Design, held in Montpellier, France, from 1 - 2 June 1995. The workshop aimed to provide a forum for researchers and practitioners from a variety of backgrounds to discuss the many facets of hypermedia design. Among the specific topics covered by the papers are: design methods, multimedia modelling, higher structures in hypermedia design spaces, user-interface design for hypermedia, building distributed web applications, and hyperdialogs. The resulting volume provides a comprehensive overview of the state of the art in this important field. It will be of interest to researchers, practitioners and students involved in any aspect of hypermedia design.