Bernard Bekavac, Marc Rittberger Information Retrieval in a Regional, Distributed Information Area The 'Electronic Mall Bodensee' (EMB) is a distributed information space for the Lake Constance Area. The main goal is to offer all kind of information suitable for companies administration and the population in this area along an electronic distributed system everybody can easily achieve and handle. The World Wide Web fulfills the mentioned requirements and is used as the basic concept to build the EMB. Besides the mostly used HTML pages at the EMB one can find other types of resources in the EMB, as ftp servers, sql based databases, other textuel non standard databases, VRML based data. Other types of sources are planned to be integrated into the EMB. We want to introduce a system in this regional information space to identify documents or information spaces relevant to a users query. To achieve this goal one needs to have information about the information resources available, an effective and efficient search engine and a data fusion mechanism to bring together the relevant documents from different information sources. Information resources: To identify information resources there needs to be any kind of description or classification of the source, automatically generated by the source itself like an index or written by a human or a machine. In the EMB environment any EMB server will inform a central search server about the data it stores. The information may be the index or a formal description or another representation. A daemon on every EMB server will announce any changes on its server to the central search server handling the resource selection. The information available about information sources wil be handled in a hypertext environment by its own. Search engine: Besides the classification or indexing of data every EMB server needs to inform the EMB central search server about the structure of a EMB hypertext on a specific server. This structure will be presented to the EMB central search server. The structure information will be used besides the content information to decide which EMB server and which information will be searched and presented to the user. Data fusion: Any given documents from the EMB information space will be arranged in a hypertext environment to present not only a list of relevant pages, but a list of relevant areas to look for information. These areas will be identified by the structure information passed to the central server. This server uses the structure information besides the content information of the information units in the system to create its own information structure, where a user can look up for relevant information. Therefore a list of some small, structured information spaces will be presented to the user. Main emphasis in the project are the integration of hypertext structures into the search process and the communication between the user and the system to identify resources and to interact with the results presented by the system.