Interactive ambient visualizations for soft advice

Some software packages offer the user soft advice: recommendations that are intended to help the user create high quality artifacts, but which may turn out to be bad advice. It is left to the user to determine whether the soft advice really will improve quality, and to decide whether or not to adopt it. Visualizations can help the user in making this decision, but we believe that conventional visualizations are less than ideal. In this paper, we describe an interactive ambient visualization to help users identify, understand and interpret soft advice. Our visualization was developed to help programmers interpret code smells, which are indications that a software project may be suffering from design problems. We describe a laboratory experiment with 12 programmers that tests several hypotheses about our visualization. The findings suggest that our tool helps programmers to identify smells more effectively, and to make more informed judgments about the design of the program under development. We then describe an application of our visualization technique in another domain: an English style and grammar advisor. This second application suggests that our technique can be applied to several domains, and also suggests how the technique must be varied to make it domain specific.