Personal tools
You are here: Home → Research → Zorro

Zorro

A system for automated inference of test-driven design behaviors. (2003-2007)

Zorro

Participants
  • CSDL: Hongbing Kou
  • Affiliates: NSF

Summary

 Zorro is a Hackystat-based application that is designed to help development groups gain insight into their use of test driven design practices and the impact of this use on their software development process and products. Zorro works by analyzing process data gathered by a Hackystat sensor attached to an interactive development environment such as Eclipse. Examples of such process data includes the editing of a a test case, the compilation and building of the software, the running of a test case, and so forth.

Zorro implements a set of rules that first partitions the stream of observed developer behaviors gathered by the sensor into a sequence of "episodes", each of which may or may not correspond to a TDD cycle. A second set of rules are then applied to each episode in order to classify it as an instance of TDD or not. A pilot validation study conducted in the Spring of 2006 indicated that Zorro recognized episodes of TDD with almost 90% accuracy. Further classroom and industrial case studies provide additional evidence for Zorro's ability to support TDD recognition.

Software

Available as part of Version 7 of Hackystat.  Porting to Version 8 is in progress.

Publications

See the Zorro Publications Area for a list of publications and technical reports.

Status

Complete.

Keywords

Test driven design, empirical software engineering.
Document Actions