Current Research Information as Part of Digital Libraries and the Heterogeneity Problem

Abstract

Users of scientific information are now faced with a highly decentralized, heterogeneous document base with varied content analysis methods. Traditional providers of information such as libraries or information centers have been increasingly joined by scientists themselves, who are developing independent services of varying scope, relevance and type of development in the WWW. Theoretically, groups that have gathered current research information (CRI), literature or factual information on specialized subjects can emerge anywhere in the world. One consequence of this is the presence of various inconsistencies:

- Relevant, quality-controlled data can be found right next to irrelevant and perhaps demonstrably erroneous data.

- In a system of this kind, descriptor A can assume the most disparate meanings. Even in the narrower context of specialized information, descriptor A, which has been extracted in an intellectually and qualitatively correct manner, and with much care and attention, from a highly relevant document, is not to be compared with a term A that has been provided by automatic indexing in some peripheral area.

Thus, the main problem to be solved is as follows: users must be supplied with heterogeneous data from different sources, modalities and content analysis processes via a visual user interface without inconsistencies in content analysis, for example, seriously impairing the quality of the search results. A scientist who, for example, is looking for social science information on subject X does not first want to search the social science literature database SOLIS and the current research database FORIS, and then the library catalogues of the special compilation area of social sciences at the library catalogues and in the WWW – each time using different search strategies. He wants to phrase his search query only once in the terminology to which he is accustomed without dealing with the remaining problems.

Closer analysis of this problems shows that narrow technological concepts, even if they are undoubtedly necessary, are not sufficient on their own. They must be supplemented by new conceptual considerations relating to the treatment of breaks in consistency between the different processes of content analysis. Acceptable solutions are only obtained when both aspects are combined. The IZ research group (Bonn, Germany) is working on this aspect in four different projects: Carmen, ViBSoz, ELVIRA and the ETB project. Initial solutions for transfer modules are available now and will be discussed.