The MINSTREL project has identified and developed software techniques for future office filing and retrieval systems. The focal point was the development of an office information model which could be used to describe the properties of and operations used on all forms of office information.
In MINSTREL, the existence of a large and varied body of information composed of images, sound, graphics, text and numerical data was assumed, together with a demand for greater functionality and a higher degree of integration in computerised office support tools. Consequently, it was felt that future office systems should be built with a more flexible and powerful architecture. The chosen architecture was based on three major principles: modularisation, layering of functions, and a single uniform data representation.
The central task was the development of an office information model which could provide a single uniform data representation, and thus play a role similar to a database management system. The other tasks in MINSTREL were aimed at developing techniques to improve the effectiveness with which an office worker can retrieve required information from such a filing system. The ideas developed in the project were validated by integrating them in a prototype implementation of an office filing system incorporatingfeatures, including a graphic-based user interface and efficient storage structures.
The following results were achieved:
-a complete formal specification of a comprehensive office information model
-a prototype implementation of a subset of this office information model, including a specially developed storage management system for efficient access
-a prototype implementation of a dialogue manager providing a uniform user interface to all office applications
-the use of graphics to construct a user-friendly interface employing various presentation forms to clearly manifest the underlying organisation and structure of the stored data
-a formal specification and prototype of a query language that allows convenient querying of both structured data and text, and which can also handle imprecision, both in query criteria and in the stored data values
-a text retrieval algorithm which uses syntactical analysis to achieve a more accurate match of queries and documents
-the specification of an access control system, within the office information model, which provides a pseudo-mandatory as well as a discretionary security policy
-a prototype implementation of a data entry subsystem that can input paper documents using a scanner in an adaptive manner, in that it can separate images from text and recognise characters despite poor-quality copies, font changes and other irregulariti es.
Some of the results, such as the graphic user interface, the handling of imprecision, the content retrieval mechanism, the access control system and the data-entry subsystem, can be applied to present-day office information systems. Some of these will be exploited by the industrial partners. The system architecture and the office information models are very relevant for future office information systems with major requirements for multimedia data and the integration of application programs. In particular,the office information model is ideally suited for representing structured documents, as in the Office Document Architecture (ODA) standard, and for integrating them into a general filing and retrieval system.
X Dublin 2
X Dublin 4