Andrew Dinn

Dimensions of Active Behaviour

This paper introduces a number of dimensions of active rule system behaviour which can be used both to highlight differences between proposals for active rule systems, and to identify the requirements of different applications. These dimensions relate to the structure, execution model and management of active rules, and enable the concise expression of what facilities a system supports and what features an application requires.

Design and implementation of ROCK & ROLL: a deductive object-oriented database system

This paper presents an approach to the development of a deductive object-oriented database system, describing the key design decisions and their consequences for implementation. The approach is novel, in that it integrates an object-oriented database system manipulated using an imperative programming language (ROCK) with a logic language for expressing queries and methods (ROLL). The integration is made seamless by deriving both the imperative and logic languages from a single formally defined data model, thereby avoiding impedance mismatches when they are integrated.

Geographic Data Handling in a Deductive Object-Oriented Database

This paper describes how a deductive object-oriented database (DOOD) can be used to support the storage and management of data which is typical of that found in geographic information systems (GIS). This is done with two aims in mind: to illustrate how a combination of deductive and object-oriented facilities can be applied effectively in an advanced application, thereby motivating the development of DOOD systems; and to show how geographic database systems stand to gain from the utilisation of advanced data modelling and inference facilities as supported by a DOOD. The paper describes the DOOD system which has been used for prototyping a range of geographic concepts, presents a framework for the structural organisation of GIS data using an object-oriented data model, and shows how a logic query language can be used within this structural framework to perform a range of analyses.

VESPA: A Benchmark for Vector Spatial Databases

Facilities for the storage and analysis of large quantities of spatial data are important to many applications, and are central to geographic information systems. This has given rise to a range of proposals for spatial data models and software architectures that allow database systems to be used cleanly and efficiently with spatial data. However, although many spatial database systems have been built, there have been few systematic comparisons of the functionality or the performance of such systems. This is probably at least partly due to the lack of a widely used, standard spatial database benchmark. This paper presents a benchmark for vector spatial databases that covers a range of typical GIS functions, and shows how the benchmark has been implemented in two systems: the object-relational database PostgreSQL, and the deductive object-oriented database ROCK & ROLL extended to support the ROSE algebra. The benchmark serves both to evaluate the facilities provided by the systems and to allow conclusions to be drawn on the efficiency of their spatial storage managers.

An Active Rule Language for ROCK & ROLL

This paper presents an active rule language for the ROCK & ROLL deductive object-oriented database system. A characteristic feature of ROCK & ROLL is that it blends imperative and deductive programming styles so that both can be used together in support of passive database applications. The aim in developing an active extension is to allow declarative expression of aspects of active behaviour wherever possible, without imposing prohibitive restrictions on the power of the resulting system. The proposal which results is more powerful than most earlier declarative active rule systems, in both its language and execution model, without resorting to the wholly procedural approach supported by most proposals for active object-oriented databases. The paper indicates where retaining declarative features yields greatest benefits, but also where difficulties are encountered which lead to compromises.

Active rule analysis and optimisation in the rock & roll deductive object-oriented database

Abstract Active database systems provide facilities that monitor and respond to changes of relevance to the database. Active behaviour is normally described using Event Condition Action rules (ECA-rules), and a number of systems have been developed, based upon different data models, that support such rules. However, experience using active databases shows that while such systems are powerful and potentially useful in many applications, they are hard to program and liable to exhibit poor performance at runtime. This document addresses both of these issues by examining both analysis and optimisation of active rules in the context of a powerful active database system. It is shown how rule analysis methods developed for straightforward active rule languages for relational data models can be extended to take account of rich event description languages and more powerful execution models. It is also shown how the results of analyses can be exploited by rule optimisers, and that multiple query optimisation can be applied in a range of circumstances to eliminate duplicated work during rule processing.

The Implementation of a Deductive Query Language over an OODB

The ROCK & ROLL database system cleanly integrates deductive and object-oriented capabilities by defining an imperative programming language, ROCK, and a declarative, deductive language, ROLL, over a common object-oriented (OO) data model. Existing techniques for evaluation and optimization of deductive languages fail to address key requirements imposed by ROLL such as: strict typing; placement of deductive methods (predicates) within classes; encapsulation; overriding and late binding. This paper describes the task of implementing an evaluator and optimizer for ROLL, explaining how existing implementation techniques for deductive languages were adapted to meet these requirements and extended to support novel types of optimization.

Extending a deductive object-oriented database system with spatial data handling facilities

Abstract This paper describes the integration of a spatial data-handling component with the ROCK & ROLL deductive object-oriented database system. The extended ROCK & ROLL system provides much more comprehensive and better integrated database programming facilities than other candidate platforms for spatial information systems. The extended system serves developers with an intuitive, expressive, formally defined collection of spatial data types as primitive types whose operations have state-of-the-art computational complexity. The integration of these types with the object-oriented modelling, imperative programming and deductive querying facilities of ROCK & ROLL makes available a comprehensive and integrated suite of complementary mechanisms for the development of spatial information systems. The paper also provides preliminary benchmark results which indicate that kernel-support for spatial data handling does yield performance gains and that the extended ROCK & ROLL system compares well with a specialist geographic information system and two widely known extensible database systems when the latter are extended with spatial data handling facilities.

RAP: The ROCK & ROLL Active Programming System

This chapter describes RAP (ROCK & ROLL Active Programming system), an active rule system embedded within the ROCK & ROLL deductive object-oriented database system. A brief description of ROCK & ROLL is followed by an overview of RAP and a categorization of its Knowledge and Execution Models according to the dimensions introduced in Chapter 1.