Workflow data footprints

Traditional business workflows are activity-centric, i.e., they focus mostly on specifying the control-flow aspect of business processes. The data aspect, although equally relevant, is attached to the control-flow later, as a second rate citizen, making the actual document sharing and exchange much less transparent and harder to verify. In this paper we introduce the notion of a data footprint, a directed graph representing the flow of some (user-specified) data elements, abstracting, as much as possible, from all the other aspects. We show how data footprints can be obtained automatically and efficiently from every activity-centric workflow that includes some conceptual data information. We also show how a footprint can be used to discover data-flow errors, or to compare the modeled object life-cycle with the required one