Dynamic dimensional feedback: An interface aid to business rule creation

Abstract

We examined the efficacy of providing dimensional feedback in the user interface as people construct business rules. Business rules often involve objects that have dimensions (e.g., area, cost, weight) and may entail complex calculations on these objects. Such mathematical expressions are error prone. We designed and tested a novel interface utilizing dimensional analysis to provide advice on expected dimensions, and error feedback on incorrect usage of dimensional objects. Experimental studies were carried out in which subjects used the interface to create rules based on word problems. In a balanced design, rule creation with dimensional feedback was compared to rule creation without such feedback. We found evidence for the usefulness of such feedback. In addition we observed the need to support higher level 'proxy variables' and stepwise definition of complex rules. The findings have implications beyond business rules tools and may be applied to any system requiring mathematical expressions with dimensional objects.