Topics in Information Systems – serie

Current experimental systems in industry, government, and the military take advantage of knowledge-based processing. For example, the Defense Advanced Research Projects Agency (DARPA), and the United States Geological Survey (USGS) are supporting the develop­ ment of information systems that contain diverse, vast, and growing repositories of data (e.g., vast databases storing geographic informa­ tion). These systems require powerful reasoning capabilities and pro­ cessing such as data processing, communications, and multidisciplinary of such systems will scientific analysis. The number and importance grow significantly in the near future. Many of these systems are severely limited by current knowledge base and database systems technology. Currently, knowledge-based system technology lacks the means to provide efficient and robust knowledge bases, while database system technology lacks knowledge representation and reasoning capabilities. The time has come to face the complex research problems that must be solved before we can design and implement real, large scale software systems that depend on knowledge-based processing. To date there has been little research directed at integrating knowledge base and database technologies. It is now imperative that such coordinated research be initiated and that it respond to the urgent need for a tech­ nology that will enable operational large-scale knowledge-based system applications.

The growing demand for systems of ever-increasing complexity and precision has stimulated the need for higher level concepts, tools, and techniques in every area of Computer Science. Some of these areas, in particular Artificial Intelligence, Databases, and Programming Lan­ guages, are attempting to meet this demand by defining a new, more abstract level of system description. We call this new level conceptual in recognition of its basic conceptual nature. In Artificial Intelligence, the problem of designing an expert system is seen primarily as a problem of building a knowledge base that repre­ sents knowledge about an enterprise. Consequently, Knowledge Repre­ sentation is viewed as a central issue in Artificial Intelligence research. Database design methodologies developed during the last five years are almost unanimous in offering semantic data models in terms of which the designer directly and naturally models an enterprise before proceed­ ing to a detailed logical and physical database design. In Programming Languages, different forms of abstraction which allow implementation­ independent specifications of data, functions, and control have been a major research theme for a decade. To emphasize the common goals of these three research efforts, we call this new activity conceptual modelling.

There is a growing interest in integrating databases and programming languages. In recent years the programming language community has developed new models of computation such as logic programming, object-oriented programming and functional programming, to add to the well established von Neumann model. The data base community has almost independently developed more and more sophisticated data models to solve the problems of large scale data organisation. To make use of these new models in programming languages there must be an awareness of the problems of large scale data. The data base designers can also learn much about language interfaces from programming language designers. The purpose of this book is to present the state of the art in integrating both approaches. The book evolved from the proceedings of a workshop held at the Appin in August 1985. It consists of three sections. The first, "Data Types and Persistence", discusses the issues of data abstraction in a persistent environment. Type systems, modules and binding mechanisms that are appropriate for programming in the large are proposed.Type checking for polymorphic systems and across innovations of the type checker are also discussed. The second section, "Database Types in Programming Languages", introduces the concept of inheritance as a method of polymorphic modelling. It is shown how inheritance can be used as a method of computation in logic programming and how it is appropriate for modelling large scale data in databases. The last section discusses the issues of controlled access to large scale data in a concurrent and distributed persistent environment. Finally methods of how we may implement persistence and build machine architectures for persistent data round off the book.

There is an established interest in integrating databases and programming languages. This book on Data Types and Persistence evolved from the proceedings of a workshop held at the Appin in August 1985. The purpose of the Appin workshop was to focus on these two aspects: persistence and data types, and to bring together people from various disciplines who have thought about these problems. Particular topics of"interest include the design of type systems appropriate for database work, the representation of persistent objects such as data types and modules, and the provision of orthogonal persistence and certain aspects of transactions and concurrency. The programme was broken into three sessions: morning, late afternoon and evening to allow the participants to take advantage of two beautiful days in the Scottish Highlands. The financial assistance of the Science and Engineering Research Council, the National Science Foundation and International Computers Ltd. is gratefully acknowledged. We would also like to thank Isabel Graham, Anne Donnelly and Estelle Taylor for their help in organising the workshop. Finally our thanks to Pete Bailey, Ray Carick and Dave Munro for the immense task they undertook in typesetting the book. The convergence of programming languages and databases to a coherent and consistent whole requires ideas from, and adjustment in, both intellectual camps. The first group of chapters in this book present ideas and adjustments coming from the programming language research community. This community frequently discusses types and uses them as a framework for other discussions.

This book is an anthology of the results of research and development in database query processing during the past decade. The relational model of data provided tremendous impetus for research into query processing. Since a relational query does not specify access paths to the stored data, the database management system (DBMS) must provide an intelligent query-processing subsystem which will evaluate a number of potentially efficient strategies for processing the query and select the one that optimizes a given performance measure. The degree of sophistication of this subsystem, often called the optimizer, critically affects the performance of the DBMS. Research into query processing thus started has taken off in several directions during the past decade. The emergence of research into distributed databases has enormously complicated the tasks of the optimizer. In a distributed environment, the database may be partitioned into horizontal or vertical fragments of relations. Replicas of the fragments may be stored in different sites of a network and even migrate to other sites. The measure of performance of a query in a distributed system must include the communication cost between sites. To minimize communication costs for-queries involving multiple relations across multiple sites, optimizers may also have to consider semi-join techniques.

The term "Office Automation" implies much and means little. The word "Office" is usually reserved for units in an organization that have a rather general function. They are supposed to support different activities, but it is notoriously difficult to determine what an office is supposed to do. Automation in this loose context may mean many different things. At one extreme, it is nothing more than giving people better tools than typewriters and telephones with which to do their work more efficiently and effectively. At the opposite extreme, it implies the replacement of people by machines which perform office procedures automatically. In this book we will take the approach that "Office Automation" is much more than just better tools, but falls significantly short of replacing every person in an office. It may reduce the need for clerks, it may take over some secretarial functions, and it may lessen the dependence of principals on support personnel. Office Automation will change the office environment. It will eliminate the more mundane and well understood functions and will highlight the decision-oriented activities in an office. The goal of this book is to provide some understanding of office . activities and to evaluate the potential of Office Information Systems for office procedure automation. To achieve this goal, we need to explore concepts, elaborate on techniques, and outline tools.

Information systems are large repositories of factual and inferential knowledge intended to be queried and maintained by a wide variety of users with different backgrounds and work tasks. The community of potential information system users is growing rapidly with advances in hardware and software technology that permit computer/communications support for more and more application areas. Unfortunately, it is often felt that progress in user interface technology has not quite matched that of other areas. Technical solutions such as computer graphics, natural language processing, or man-machine-man communications in office systems are not enough by themselves. They should be complemented by system features that ensure cooperative behavior of the interfaces, thus reducing the training and usage effort required for successful interaction. In analogy to a human dialog partner, we call an interface cooperative if it does not just accept user requests passively or answer them literally, but actively attempts to understand the users' intentions and to help them solve their applica­ tion problems. This leads to the central question addressed by this book: What makes an information systems interface cooperative, and how do we provide capabilities leading to cooperative interfaces? Many answers are possible. A first aspect concerns the formulation and accep­ tance of user requests. Many researchers assume that such requests should be formulated in natural language.

In the past, applied artificial intelligence systems were built with particular emphasis on general reasoning methods intended to function efficiently, even when only relatively little domain-specific knowledge was available. In other words, AI technology aimed at the processing of knowledge stored under comparatively general representation schemes. Nowadays, the focus has been redirected to the role played by specific and detailed knowledge, rather than to the reasoning methods themselves. Many new application systems are centered around knowledge bases, i. e. , they are based on large collections offacts, rules, and heuristics that cap­ ture knowledge about a specific domain of applications. Experience has shown that when used in combination with rich knowledge bases, even simple reasoning methods can be extremely effective in a wide variety of problem domains. Knowledge base construction and management will thus become the key factor in the development of viable knowledge-based ap­ plications. Knowledge Base Management Systems (KBMSs) are being proposed that provide user-friendly environments for the construction, retrieval, and manipUlation of large shared knowledge bases. In addition to deductive reasoning, KBMSs require operational characteristics such as concurrent access, integrity maintenance, error recovery, security, and perhaps distribution. For the development ofKBMSs, the need to integrate concepts and technologies from different areas, such as Artificial Intel­ ligence, Databases, and Logic, has been widely recognized. One of the central issues for KBMSs is the framework used for knowledge representation-semantic networks, frames, rules, and logics are proposed by the AI and logic communities.