Summarizing Traces as Signals in Time
Adrian Kuhn, Orla Greevy · BORIS (University Library Bern) · 2006
One of the key challenges of dynamic analysis approaches is that they imply a huge volume of data, thus making it difficult to extract high level views. In this paper we describe a novel approach to trace summarization by visually representing entire traces as signals in time. Our technique produces a visualization of the complete feature space of a system that fits on one page. The focus of our work is to visually represent individual traces feature behavior. We assume a one-to-one mapping between features and traces. We apply the approach on a case study, and discuss how our visualization supports the reverse engineer to identify patterns in traces of features. Moreover, we show how the visual analysis of our trace signals reveals that assumed one-to-one mappings between features and traces may be flawed.