OO Relative Complexity Metric

OO Relative Complexity Metric

[Jianqiang Zhuo]

Here at the Idaho Software Engineering Test Lab we have created
a modified relative complexity metric for OO code.  Unlike traditional
metric models, the modified relative complexity metric takes into
consideration both class-level and procedural-level (e.g., functions
and methods) complexity, and aggregates those complexities into a 
single value representing the complexity of an OO file, subsystem or
test functionality.  Initial tests with industrial code show that it
is promising for both fault prediction and test resource allocation.

But we need more data to validate it.  Does anyone have any OO code
or metrics that you can share with us?  We would analyze your data and
show you the results as well as the steps necessary to compute the 
new metrics.  At a minimum, what we need are:

   a) class and method level complexity metrics for OO code
   b) an a priori analysis of that code; either fault data, complexity
      rank ordering, quality circle analysis, construction or maintenance
      effort, or similar measurement.

Optionally, if you have C++ code and can share it with us, we'd be glad to
compute the metrics and do the analyses.

Please contact myself, Michael Zhuo (michael@silver.cs.uidaho.edu) or
Dr. Paul Oman (oman@cs.uidaho.edu) if you think you have data
that would work and you want to get it analyzed.

Thanks, Michael Zhuo
Software Engineering Test Lab
University of Idaho.