As the mounds of information and the number of Internet users grow, the problem of indexing
and retrieving of electronic information resources becomes more critical. The existing search
systems tend to generate misses and false hits due to the fact that they attempt to match
the speci ed search terms without proper context in the target information resource. In
environments that contain many di erent types of data, content indexing requires type-
speci c processing to extract indexing information e ectively. The COncordia INdexing and
DIscovery (Cindi) system is a system devised to support the registration of indexing meta-
data for information resources and provide a convenient system for search and discovery.
The Semantic Header, containing the semantic contents of information resources stored in
the Cindi system, provides a useful tool to facilitate the searching for documents based on a
number of commonly used criteria. This paper presents an automatic tool for the extraction
and storage of some of the meta-information in a Semantic Header and the classi cation
scheme used for generating the subject headings.