Mantis Bug Tracker - Forums

I am wondering about what the intent of the calculations of the Reporter Effectiveness Section is?

Based on looking through the code, I notice that the Severity number is based on an aggregate of the severity of a bug multiplied by a qualifying factor. More severe bugs will carry a greater weight.

# These are our overall "values" for severities and non-bug results
$t_severity_multiplier[FEATURE] = 1;
$t_severity_multiplier[TRIVIAL] = 2;
$t_severity_multiplier[TEXT] = 3;
$t_severity_multiplier[TWEAK] = 2;
$t_severity_multiplier[MINOR] = 5;
$t_severity_multiplier[MAJOR] = 8;
$t_severity_multiplier[CRASH] = 8;
$t_severity_multiplier[BLOCK] = 10;
$t_severity_multiplier['average'] = 5;

The False number is also based on a similar calculation, but for each bug reported that really is not a bug, the multiplier quoted above is multiplied by a "gotcha" number, meaning that the punishment is greater than the crime.

$t_notbug_multiplier[UNABLE_TO_DUPLICATE] = 2;
$t_notbug_multiplier[DUPLICATE] = 3;
$t_notbug_multiplier[NOT_A_BUG] = 5;

Finally, Total is the simple difference between Severity and False.


My concern is this: for a Reporter that reports few issues where many of them are duplicates or we are not able to reproduce the bug, then it is possible to have a negative effectiveness based on the summary screen. It is my opinion that this may be taken in the wrong context and a "why bother" position may be adopted to reporting bugs. I suggest that the False number should be closer to a percentage of the Severity number, for example 0.75. This way, there is credit for making the effort to communicate the bugs and a "penalty" may still be imposed if that is the desire. At this point, the code is not setup for fractional numbers and the above factors are not definable at the config_inc.php level.

I also believe that Effectiveness is best represented as a percentage, not a whole number. A calculation of this would be: Severity/Total*100.

I welcome comments on my thoughts.

Thank you.

Hi All,

Does anyone know if since 2005 (when this thread was open) there's been any change in the values and logic for calculating effectiveness?

(sorry if this was already discussed, I found this via Google search, and a brief search on the forum with keyword 'effectiveness' didn't return anything of the kind)