The integration of IAI's language technology into an existing Web-based groupware system such as the BSCW shared workspace system is challenging because it would be the first groupware system that fully leverages the power of existing language technology.

The following language technology components will be adapted to the specific "int.unity" requirements:

Term Mining (TM): Terminology is of crucial importance for the set-up of a collaborative communication and shared workspace system environment because it defines the concepts of the domain and the different language-specific realisations of these concepts including possible synonymy and antonymy. TM is based on a human language analysis module (morphology and shallow syntactic/semantic parsing; aka IAI core language technology) and delivers possible term candidates together with frequency information and context information. The XML output is suitable for further processing in XML enabled environments.

Language Proofing (LP): Spelling and grammatical correctness is essential for communicating information in different forms. The compliance to predefined sets of domain vocabularies (terminology) and certain writing rules according to style guidelines improves the employment of language technology and permits the communication in multiple languages.

Bi-lingual Glossary (BG): The selection of the right wording ensures smooth communication and translation between the "int.unity" community members. The generation of a bi-lingual word list supports this selection.

Bi-lingual Search (BS): The search for specific information in different languages is useful to avoid unnecessary translation. BS supports this demand for cross-language information retrieval.

Information Gisting: The raw translation of an information object instance from a foreign language into the native language and vice versa permits the assessment of the object instance's relevance for a set of given information search parameters. The compliance with a given multi-lingual terminology ensures that the right information is preserved during the translation process.

Summarisation (SUM): The extraction of the important facts of an information object instance, mono- and multi-lingual, permits better and easier navigation within a given information space, and information filtering according to specific personal needs and demands.


For the adaptation of the IAI language technology the following steps have to be implemented:

language technology Resource Building: Term mining ensures the selection and building of the appropriate bi-lingual terminology repository of "int.unity" and forms the vocabulary basis for the adaptation of the other components. For this "int.unity" information objects are processed and the user community is responsible for the selection of the relevant terminology. The result of this terminology selection will be converted into the language technology internal format of the IAI core language technology and stored in a term base with an "int.unity"-specific user interface.

language technology Rule Base Building: Language proofing ensures that the communication will be fulfilled according to a set of well-defined "language rules". The set of core rules for English and German has to be enlarged according to the style used in the information objects of the "int.unity" consortium. For the trial only a subset of these information objects can be assessed and validated. The identified "int.unity"-specific grammar and style rules will be checked for implementability, i.e. formalisation, and then integrated into the existing software and lingware components. Since this is accomplished step by step, there are several testing phases through the users until a stable "int.unity" configuration of the language proofing is implemented.

language technology Integration: To fully leverage the power of the adapted language technology the new language technology components have to be integrated into the overall "int.unity" environment.