Skip to main content

Intelligent Workstation

Objective

This project aimed to develop a complete system to help office workers do their job. In particular, office activities were modelled (documents, organisation and procedures) and these representations used to provide advice on actions required. Hardware andsoftware aspects were concerned with application-oriented developments and the user interface. For this purpose the project was divided into six modules:
-The assembly of the hardware for the construction of the workstation. Dedicated hardware, such as the filtering mechanisms developed during the first year, were taken into account in assembling the workstation and servers. The hardware was selected from products available on the market.
-The operating system of the workstation. This was network-oriented and supported multitasking and synchronisation with all the necessary mechanisms for increasing the reliability of the distributed system.
-The implementation of the LISP environment and Knowledge Representation System (KRS).
-The layer of office system tools that embed office semantics.
-Concrete applications implemented with the office system tools.
-The improvement of the user interface. This module was divided into three parts: natural language processing and an author system; enhancing the user interface with multimedia features (sound and image); and defining and implementing a conceptual modelof the user interface.
The project used artificial intelligence (AI) techniques to design an office system that employed both knowledge about office organization and procedures, and the necessary tools (natural language and multimedia) for an easy to use interface with the user. The workstation hardware, based on the BULL Metaviseur provided a table top office engine with a large, fast access memory capacity, efficient graphics and voice support, a 32-bit microprocessor, Ethernet communications support and a telephone interface.
The operating system environment was a Unix-based and had efficient graphic and audio support. The intelligent environment is made up of a LE-LISP compiler with its object oriented and virtual graphics extensions. A set of natural language processing (NLP) tools were built for English and Dutch. A dialogue system implemented above the knowledge representation system (KRS) used these tools to enable natural language query expressions onthe office model and natural language answer formulations.
The KRS was tested and extended to include the fast construction of knowledge based applications. The user interface management system (UIMS) was developed as a tool for assisting the easy design and reimplementation of office application user interfaces. This is related to the development of ACHILLES, a general tool for the faster specification and generation of user interfaces for particular applications.
The project used AI techniques to design an office system that employed both knowledge about office organisation and procedures, and the necessary tools (natural language and multimedia) for an easy-to-use interface with the user The basic components and tools for complex applications were successfully developed and verified. Interim versions of several prototypes were built and demonstrated at the Hannover Fair and the ESPRIT Conference Week.
The workstation hardware, based on the BULL Metaviseur, provides a table-top office engine with a large, fast-access memory capacity (40-190 Mbytes on a 5.25" hard disk), efficient graphics and voice support, a 32-bit microprocessor, Ethernet communications support and a telephone interface.
The operating system environment is Unix-based and has efficient graphic and audio support. The intelligent environment is made up of a LE-LISP compiler with its object-oriented and virtual graphics extensions.
A set of Natural Language Processing (NLP) tools were built for English and Dutch. A dialogue system implemented above the KRS uses these tools to enable natural language query expressions on the office model and natural language answer formulations.The KRS was tested and extended to include the fast construction of knowledge-based applications. KRS has been proved to be ideally suited for research laboratories investigating advanced knowledge-representation issues. At the same time it can also be used as a knowledge engineering environment for the construction of complex interfaces, or for other knowledge-based tasks.
The User Interface Management System (UIMS) was developed as a tool for assisting the easy design and reimplementation of office application user interfaces. This is related to the development of ACHILLES, a general tool for the faster specification and generation of user interfaces for particular applications.
Exploitation
A complete prototype of an office workstation connected to servers through a network and based on artificial intelligence techniques is available. The results will be immediately transferable into a broad spectrum of workstations. These will be the focal point of information systems in the 1990s. Spin-offs from the IWS have already been introduced in product developments.
In the areas of specific exploitation:
-Bull are developing a programmatic interface for the X.400 mail system used in a multimedia server. A graphics-oriented browsing tool derived from GSE will be used in conjunction with the server. Bull is also expanding its activities in X-Windows by act ively participating in X-Windows development activities within the scope of OSF.
-The ACHILLES system has been used in the large MUSE project funded by NATO, assisting the development and reimplementation of the MUSE system's highly complicated user interface. This work has been undertaken by Forth.
-Oc and the Katholieke Universiteit Nijmegen both participated in a national project on natural language processing, Stimuerings Project Informatica Nederland (SPIN).
-A spin-off company (Knowledge Technologies CV) has been set up in Italy for the commercialisation of KRS and to contribute to KRS in southern Europe.

Coordinator

BULL SA
Address
7 Rue Ampère
91343 Massy
France

Participants (5)

FORTH-RESEARCH CENTER OF CRETE
Greece
Address
Daidalou P.o.box 1385
10 Iraklion
Institut National de Recherche en Informatique et en Automatique - INRIA
France
Address
Domaine De Voluceau-rocquencourt
78153 Le Chesnay
KATHOLIEKE UNIVERSITEIT NIJMEGEN
Netherlands
Address
Montessorilaan 9104
6500 HE Nijmegen
OCE-NEDERLAND BV
Netherlands
Address
St. Urbanusstraat Postbus 101
5900 MA Venlo
VRIJE UNIVERSITEIT BRUSSEL
Belgium
Address
Pleinlaan
1050 Brussels