Exploring design principles of task elicitation systems for unrestricted natural language documents

During the design of interactive systems, user tasks need to be identified within natural language documents (like interview transcripts, support messages or workshop memos) and be transformed into task models. This time-consuming and error-prone analysis process demands for automation, however corresponding software support is still sparse. This paper describes a Design Science Research project, which explores design principles for a system aiming to close this gap. To evaluate the principles, they are instantiated in an innovative artifact called REMINER which combines Information Retrieval, Natural Language Processing and Annotation technology. The artifact can be used to semi-automatically identify user tasks from unrestricted natural language documents and to organize them into task models. Results of two extensive evaluations of the artifact show, that it considerably addresses the underlying problem areas of this process.