Practices that have mediocre relationship to reduced escaped defects. Several of the practices with mediocre relationship are either deployed ineffectively or effect other software quality practices other than reliability. Frestimate's "mediocre software practice analyzer" assesses the percentage of all software development practices that aren't reducing software defects.
Mediocre software development practice analyzer
The foundation practices are those that nearly every project in our database had in place. These foundation practices should not be skimped as they are prerequisites for several other practices that effect escaped defects and reliability.
Foundation software development practices analyzer
Having the right development practices is in place - but only if the related prerequisite practices are in place. Software organizations have a tendency to choose practices that are "interesting" while neglecting the basic practices that are less interesting but essential. The development practices that you identify in the prediction surveys are stacked based on the skill level typically required for each practice. The practices that are employed at the 90 percentile group can typically be employed by nearly all software organizations while the practices associated with the 3 or 10 percentile groups may require too many prerequisites for many organizations to employ.
If your profile is skewed to the left then your practices have been implemented in relatively the correct order with basic practices coming first. If your profile is skewed to the right (such as with the full-scale survey in blue shown below),, your organization is attempting to employ practices that require advanced capabilities while ignoring the practices that are foundational. The sensitivity summary graphs the responses to all three of the Full-scale survey results.
Software reliability sensitivity analysis graph
The sensitivity analysis tab also allows you to estimate the staffing for the released software as well as the staffing for the testing of the software. According to our research, the most common reason for a software project to be late is that the people
schedule to work on the release are sidetracked by unplanned maintenance on previous releases. Predicting the maintenance staffing can also be used to determine warranties.
Use software reliability to predict software maintenance staffing
Use software reliability to predict software testing staffing