A task and pattern-based modeling approach for knowledge sharing systems

Motivation -- Adaptable systems need flexible design strategies. Our approach combines task- and pattern-based a priori modeling for the construction of context and situation aware target systems in the knowledge sharing domain. Research approach -- An extensive study, the p.i.t.c.h. project [3], [5], [6] with partners from industry and research was conducted in order to define a hierarchical pattern language with domain, structure, HCI and transformation patterns, find candidate patterns for each category and map individual knowledge sharing task models to these patterns. Findings/Design -- The results show that tasks models offer a flexible way to select patterns of all categories for diverse target contexts and situations. The approach also provides directions for automated generation of the target user interface. Research limitations/Implications -- Currently the pattern selection process is controlled by the developer. Future versions will try to automate the selection process in order to allow for runtime selection and generation of situation aware target user interfaces. Originality/Value -- Our approach contributes to model-based design by offering a straight-forward method for mapping task-models to patterns of various categories that can be exploited for UI generation. Take away message -- Define a broad collection of patterns, structure them hierarchically, map individual task-models to selected patterns.