What do people use for code analysis.

[Mike Brenner]

On the one hand, I would like better graphical code
analysis tools. On the other hand, I would
like to mention a different kind of code analysis.

A "Return on Investment" (ROI) philosophy would
analyze the code by how much it costs to change the code.

To do this, you can add up all the costs.

The main cost pays for analyzing the dataflows
to determine the impact of the change.

The secondary costs pays for the backlogs,
the things that you put off paying for in the past:

- the number of branches in the code not covered by automatic tests,
- the number of API parameters without tested examples,
- the dataflows without documented preconditions and postconditions,
- the hours of developer overtime not being paid,
- the cost of opening up the configuration management to a change,
- the number of platform-dependent features,
- the number of reviewer comments not kept, etc.

Notice that such things as code size, inheritance,
multiple tasks, calls, etc., don't add cost!

Code costs almost nothing compared to the true value,
the data and the dataflows.

Mike Brenner



Steven wrote:
 > Lots of code, calls to, calls by, inheritance, multiple tasks, etc.
 > What do people use to figure out what's happening?