TWENTY-ONE:
Disclosure and Dissemination of Multi-media Documents on Sustainable Development
Ir. J. ten Hagen
IE-2108 Twenty-One
Bijlmerdreef 100, P.O. box 22678, 1100 DD Amsterdam, the Netherlands
e-mail: J.tenHagen@Getronics.nl
OVERVIEW
The Twenty-One project brings together environmental organisations, technology providers and research institutes from several European countries to make documents on environmental issues - in particular on the subject of sustainable development - available on CD-ROM and on the Internet. These documents exist in a number of different media (paper, electronic documents, audio-visual materials), formats (HTML, word processed, desktop published), and in different languages. Many are not available through standard publication channels.
This diversity of media, formats and languages impedes the distribution of documents through normal channels, and makes it hard to search for and retrieve targeted material on specified subjects. The Twenty-One project ensures all these documents are made available on CD-ROM and on the Internet. The project has developed search engines that can locate the required information, and uses automatic translation tools to make foreign-language texts available. Authors and editors gain an economic advantage by the increased distribution of their documents, while users find documents more readable and more easily accessible.
Under Twenty-One:
- Paper documents are scanned and converted into electronic format;
- The document structure is analysed automatic, and the text content disclosed using OCR (Optical Character Recognition) and linguistic processing;
- Fuzzy matching is used during searching to make the system robust against spelling errors in documents and queries, and to compensate OCR errors;
- Linguistic analysis of documents makes it possible to search for relevant phrases (e.g. sustainable development), not just for individual words;
- The application of cross-language information retrieval allows for flexible searching in four languages, namely: French, English, Dutch and German.
OBJECTIVES
The main objective of Twenty-One is to develop domain-independent technology that improves the quality of electronic and non-electronic multi-media information, and which makes it more readily and cheaply accessible to a larger group of people. At the outset, Twenty-One will prove this technology in the field of ecology and sustainable development. However, the generic characteristics of the distinct software modules allow for its easy application outside the domain of environmental information exchange.
The technology developed by Twenty-One facilitates access to information by readers who are not native speakers of the language in which the information is provided. The project result is a software demonstrator that enables users to enter queries in one of four selected European languages, specifically Dutch, English, French and German, and to retrieve (multi-media) documents. The core of the system consists of an index in four languages that has been built and automatically translated by the software. The envisaged demonstrator further allows users to communicate interactively with the providers of the documents. The demonstrator uses both CD-ROM and the Internet, and its de facto applications (such as WWW, E-mail, etc.). The periodically distributed CD-ROM is used for rapid access to static document bases, whereas Internet will be used for dynamic data and document bases. In addition, the Internet will be used for interactive communication with all parties involved in the dissemination and the transaction model.
The Twenty-One information transaction model, also called the Galilei model, forms an important prerequisite to employing the technology developed within the project. The information transaction model allows different environmental organisations to exchange information, that is, to publish and retrieve different information objects. As such, the information transaction model can be viewed as a socio-technical system. This term originates from organisation theory and refers to systems where emphasis is laid on the inter-dependencies between technical equipment and the (groups of) people using this equipment.
Although the information transaction model constitutes an important framework for the success of the Twenty-One project, the main emphasis in this paper lies on the technology designed for efficient and effective document disclosure and retrieval.
PROJECT ORGANISATION, CONTEXT AND TIME SCHEDULE
Twenty-One is a project funded under the EU Telematics Applications Programme, Information Engineering sector. The project partners are from academic circles (including the University of Twente and the University of Tubingen), and commercial software companies, namely Highland Software Systems (UK) and Getronics Software (NL). In addition, Twenty-One includes sub-contract research organisations, namely DFKI (D), Rank Xerox Research Center (F), and TNO-TPD (NL), together with a number of non-profit environmental organisations, Environ Trust (UK), Friends of the Earth (B), Klima-Bündnis (D), MOOI Foundation (NL), VODO (B). The name 'Twenty-One' refers to Agenda 21, the document resulting from the UN conference on sustainable development in Rio de Janeiro in 1992. The duration of the project is 36 months with the starting date being March 1996. According to plan, the project will launch its main software products in July 1998. As a consequence, this paper reflects the work in progress.
TECHNICAL AIMS AND ACHIEVEMENTS
This section contains an overview of the technical aims of Twenty-One project. While this paper largely reflects work in progress, some modules have been designed and implemented under previous projects, and therefore it is already possible to point out some accomplishments.
The main technical product of the Twenty-One project is the so-called Twenty-One Demonstrator (to be delivered by project close). The basic functionality of the Twenty-One Demonstrator allows: (a) end-users easy and cross-language access to a multi-lingual and multi-media information base, and: (b) publishers to submit and disclose their information at very low costs.
The Twenty-One Demonstrator includes two crucial sets of software that:
- Discloses multi-media information;
- Retrieves multi-media information (accessible with current state-of-the-art Internet browsing applications, such as Netscape) from remote servers or from a local CD-ROM. The core of the Twenty-One retrieval software consists of a search kernel supporting several query modes and interface languages.
The Demonstrator's functions will be the disclosure, maintenance and retrieval of multi-media information. Disclosure represents the process of the automated attachment of features to information objects so that they can be easily found. These features in many cases will consist of index terms. Maintenance is the set of procedures that keeps the database consistent and up-to-date. This topic is not further addressed in this paper. Retrieval refers to the functionality of the system to find relevant information on the basis of a user's queries.
Design Guidelines and Principles
At the outset of the project, a number of design decisions were taken which should facilitate the development of the project software and minimise budgetary risks. As such, the following design decisions constitute a framework within which the designers, engineers and representatives of the user group agreed to cooperate.
- Reuse and extension of existing advanced technologies;
- A generic and modular, component-based architecture;
- Use of open platforms and de facto standards;
- Incremental and cooperative design;
- Evaluation planning and quality assurance.
Global System Characteristics
The aforementioned design assumptions constitute the basis for the development of the Twenty-One Demonstrator, embracing disclosure as well as retrieval functionality. The global functionality of the modules can be summarised as follows:
- Multi-media handling;
- Document conversion;
- Advanced disclosure techniques;
- Domain-tuning: Sustainable Development;
- Targeted at various publishing media;
- Cross-language retrieval/multi-linguality;
- Automatic hyperlinking.
RELATED TECHNOLOGIES, EXPLOITATION AND FUTURE PLANS
The technology with Twenty-One has been designed for efficient and effective document disclosure and retrieval. Further enhancement of this "product" could be achieved by relying on emerging technologies, such as image recognition, multi-dimensional graphic display of information spaces based on 3-D technology, and so-called intelligent agent technology. The Twenty-One technology is widely applicable. The software packages Twenty-One Publisher and Twenty-One Retrieval can assist all kinds of knowledge workers with their information needs. Examples of applications are electronic publishing (through the WWW), document information management, and workflow management. Possible user groups are, among others, publishers, researchers, marketing managers, and private information consumers. The bottom-line in the exploitation of the Twenty-One technology will be the demonstration of its problem-solving potential to any audience, but with Twenty-One technology, stakeholders should be able to work better, or reduce the 'time-to-market' for information products.
CONCLUSION
In this paper, an overview has been given of the Twenty-One approach towards electronic publishing via the Internet and/or CD ROM. By not relying on poor quality indexing, but by carefully coupling state-of-the-art multi-media processing, human language technology and various search techniques, a much more adequate environment for the dissemination of information can be realised. The employed technology seems highly supportive for the needs of the target user groups, and it is to be expected that, as a result of its use within the domain of sustainable development, it will fulfil a crucial role in triggering both publisher and end-users, in this and similar domains, to use it. Both in terms of quantity, and in terms of quality, Twenty-One will contribute to a higher level of information exchange.
| Sponsor |
European Commission, Information Engineering Sector of the Telematics Application Programme |
| Duration |
36 months, 1996-1999 |
| Partners |
Getronics Software (NL), DFKI (D), Environ Trust (UK), Highland Software (UK), Friends of the Earth (B), Klima-Bundnis (D), Rank Xerox Research Centre (F), MOOI Foundation (NL), TNO-TPD (NL), University Twente (NL), University Tuebingen (D), VODO (B). |
| WWW |
http://twentyone.tpd.tno.nl/ |
REC * PROGRAMS * TELEMATICS * DETERMINE
