TEST METRICS AND MEASUREMENTS OVERVIEW: Derive information from raw data to help in decision making Metrics derived from measurements using formulae/calculations Metrics program to decide what measurements are important and collect data Metrics and analysis helps in preventing the defects Metrics used in resource management To make right analysis on data measured To metrics related to testing and product To plan the metric for taking necessary action To describe metrics for combining the data points The metrics can be divided as 4 major units, Project metric Progress metric Productivity metric Release metric
1. PROJECT METRICS Starts with requirement gathering and end with product release Project scope get translated into size estimates Project metric can be dived as, a) Effort variance(planned VS Actual): Good idea to plot effort numbers phase-wise Calculating effort variance for each phases given Indicates effort stability b) Schedule variance(planned VS Actual): Deviation of actual schedule from estimated schedule Calculated at the end of every milestone Indicates schedule stability c) Effort distribution across phases: The various phases are captured and analysed Effort percentage for testing depends on, Type of release Amount of change to existing code base Functionality To specifies the quantum work To differentiate the baselined and revised effort in effort variance To project the causes and outcomes To estimating the total effort of testing To produce good quality product
PROBLEM: If the actual effort for effort variance is 110 and estimated effort is 100 means then, the effort variance can be calculated as, = [(110-100/100) *100 = 10% 2. PROGRESS METRICS Reflects the defects of the product Testing progress is estimated by test execution status and outcome Progress metrics can be categorized as, a) Test defect metrics: Helps to understand how defects found and assign defect priority To find defects early in test cycle using defect find rate Development to fix defects as soon as they arrive Provide a combined perspective of defects b) Development defect metrics: Helps to improving development activities Map defects to different components of product Unit of defects can be calculate using KLOC defects have been waiting to fix for a long time age analysis To reduce the defects rate To estimate the priority for defects To identifies the problem in software development model To compute the complexity and time in testing defect
PROBLEMS: If the no. of defects identified is 30 means and the size of requirement is 5,then defect density can be calculated as, DEFECT DESITY=No. of defects identified/size. = (30/5) = 6 3. PRODUCTIVITY METRICS Combines several measurements and parameters with effort on product Find out the capability of terms and purposes Describe the reason for variation in result a) Defects per 100 hours of testing: Normalize the defects Conclude that there is no end for testing b) Test case executed per 100 hours of testing: Track the productivity Judge the product quality c) Test case developed per 100 hours of testing: Modify the test cases Address new functionality and features for testing d) Defect per 100 test case: Measure the uncover defects Improve the quality e) Defect per 100 failed test case: Shows the product readiness Improve the productivity f) Test phase effectiveness: Defect found in early phases Used for planning activities
g) Close defect distribution: Judge whether all defects are fixed Analyze the closeness of defect To find some of the defects To discover the release date To estimate the quality of product PROBLEMS: If total test case executed for a period is 65and toatal hours spent in test execution is 100,then to find test cases executed per 100 hrs can be, Test cases executed per 100hrs of testing = (total test cases executed for a period/total hrs spent in test execution)*100 = (65/100)*100= 65% 4. RELEASE METRICS Determine whether the product is ready for release Decision to release product need to consider several perspective Building the credibility of metric program To increase effectiveness of metric To analyze the metric result To make decisions