Screen Shots

The prediction menu follows the same order as the Software reliability toolkit.

First, you will collect data.  You will complete one or more surveys, predict size and then predict a variety of other operational characteristics needed to convert predicted defects to predict failure rate, MTTF, reliability and availability.

The surveys include the SoftRel Shortcut model, Full-scale model, the Rome Laboratory model, your own historical data as well as the Capability lookup table and the industry lookup table.  Each of these models varies in how much time it takes to complete the survey and how much information you need to collect to do the prediction.  If you are in a hurry you can use the one parameter models such as the SEI CMMi or the industry lookup table.  The Full-scale model and Rome Labs model have many questions but are generally the most accurate.  The Shortcut model has 15 questions but has significantly more accuracy than the 1 parameter models. 

After you input the data, the results are automatically generated.

You can toggle the results to use any of the models that you have purchased.  You can also toggle the results based on defect criticality.  Results based on all defects predicted that will be serious enough to require a corrective action and results based on only those defects that will impact availability and have no workaround.

You can view the predictions for every month after delivery.  Below is the predicted failure rate for every month after deployment until the next major release.   You can see the prediction, upper and lower bounds and the filtering by failure type.

Once the results are generated you have the ability to compare your prediction to others in our database with similar project characteristics.  This is useful for sanity checking your prediction against actual completed software projects that are most similar to yours.

Now you can manage the software reliability by performing tradeoffs with cost and improvement and by generating useful management trends.

If you have the Professional Edition you can create and compare cost scenarios to determine the cheapest and fastest way to reduce the most defects. You can sort for negative survey responses in order of improvement type, correlation to lower defect density, impact on volume of reduced defect density, relative cost, relative time or a weighted score which maximizes correlation and impact and minimizes cost and time.

You can determine metrics such as the testing time needed to reduce X % of the predicted defects.

You can also graph several of the outputs such as failure rate, MTTF, unreliability and unavailability over the life of the software or until the next release.