A. Frank, M. Raubal, and M. van der Vlugt, Eds. (1999) Multi-Agency Databases to Manage Geographic Information. GeoInfo Series 16. Department of Geoinformation, Technical University Vienna, Vienna.


The co-operation between agencies forces them increasingly to share their data in intelligent ways and Multi-Agency Databases emerge. The sharing of spatial or geographic data is extremely important in modern town and regional administration, but it poses some particular and difficult to solve problems for interoperability.

A multi-agency database is composed of databases of agencies that are independent but co-operate by sharing data and have a systematic way to deal with updates originating at different locations (Müller and Frank 1999). Multi-agency databases are built on the foundation of Federated Databases, which allow data sharing and interoperability (Buehler and McKee 1996), but do not address the additional problems faced by multi-agency databases: workflow, long transaction, and transactions which do not originate with the 'owner' of the data (Chirié and Djamei 1998). The legal autonomy and responsibility of each agency hamper sharing data. Though data are shared, updating the data must be strictly controlled by the responsible individual agency. Only few officers within an agency have typically the right to update the data.

  • The problem of interoperability is that agencies have legacy systems using heterogeneous systems. These systems cannot be changed and must co-operate.
  • The problem of long transactions is that the acceptance of proposed changes by the authority of an agency can take several months, and there may be interactions between concurrent proposed updates.
  • The problem of workflow is the orderly electronic exchange of documents necessary for the administrative process managing land and other spatial resources between the multiple agencies involved.
  • The problem with updates is the processing of data changes that are initiated by other agencies but must be executed by the responsible agency.
  • COMMUTER (Co-operative multi-dimensional multimedia and topological elements reintegration) is aiming to solve these problems by using federated databases extended by a flexible transaction mechanism embedded in a workflow concept. This mechanism uses storable transaction scripts, called update proposals (Müller and Frank 1999). The workflow concepts are relevant to the multi-agency situation because they lead to distinguishing the roles of the update proposal maker initiating the proposal from the authority validating the proposal. The intended changes are stored as a transaction script listing all necessary changes. This script can be stored in a central server. The update proposals are written in a vendor-independent language, extending interoperability from the access to the data to updating it - which is novel for Geographic Information Systems (GIS).

    COMMUTER achieves interoperability with interfaces to existing data and systems. The desktop GIS software the users have already connects to a generator to produce update proposals. The legacy systems managing the databases of the agencies are interfaced with a TTML interpreter and connected to the central COMMUTER server. An email system can be used for communication between the different servers.

    The central COMMUTER server can be used as a historical database and queried by users to receive information about executed and pending update proposals, as well as update proposals in preparation. This helps orderly administration and documents how quickly the requests of citizens can be resolved.

    The COMMUTER (duration: 15th January 1998 until 31st July 1999) project was funded by the European Commission under the ESPRIT program. Major European Information Technology companies (Bull SA, Olivetti, Bull AG) and data providers (IGN, EPSILON) teamed with three public administrations (CUDL, Bologna, and Harburg) and UTE to analyse the requirements of the modern town administration. The Geoinformation Department of the Technical University Vienna joined the team to connect the project with ongoing research in GIS, databases, and interoperability.

    In this book, the solution to manage geographic data in a Multi-Agency Database is first presented in overview and then the products resulting from the project are discussed in connection with their markets. A large part of the book covers five extensive examples, where scenarios of concrete cases are described and demonstrated how they are solved. This should be helpful for the Information Managers of towns and regions to see the solutions the COMMUTER architecture can bring to them. Chapters 5 and 6 cover the two major technical results: the merging of workflow and federated database technology and the storable update proposal.

    This book shows that the COMMUTER architecture has applications beyond the environments of towns and regions, and that it opens new forms of co-operation and thus new business opportunities for agencies selling regular update services.

    Order this and/or other books from the GeoInfo Series