Page - 89 - in The Future of Software Quality Assurance

Testing in the DigitalAge 89 Fig. 3 Testing is moving from a reactive activity (test execution) through quality monitoring for operational systems towardsquality forecasting where faultsarepredicted so that theycanbefixed even before anyone notices a failure The future sees test engineers as quality forecasters that use smart sets of test cases which evolve to the best fit of addressing problem areas for an IT solution in the right variety of situations. Quality forecasting is aimed at being ahead of the test results. Before a situation occurs, digital test engineeringalready found the underlying defect. However, to get to that situation, a lot of preconditions need to be put into place. Test execution must be organized such that data is gathered in a structuredway.Qualitymonitoringmustbeorganizedinasimilarway, so thatboth testingandmonitoringdatacanbe usedas the basis forquality forecasting. Digital test engineering evolves over time. Forecasting technology coming our way helps us to be ready to find the right tools, roles, and skills that keep guarding thequalityaskedof futureproducts. 3.1 Models HelpQualityForecasting Testing is a reactive activity. Only after you test something, you know whether a product is working correctly or not. There is one constant: you never know the remainingdefects still in yourproduct. Testing must go to the situation where it can predict if and what defects will popup in thenear future.Only thencan testingkeepaheadof thequalityassurance game. Let us take weather forecast as an example. In order to predict the temperature for thecoming2weeks,50modelsarecalculated.Theyallgiveadifferentsolution. Fromthis set of results themost likelypath is chosen(byaperson!). This is something that can be done in testing as well. Assume test automation with dynamic models, generating the right test environmentsand sets of test cases using real-timedata fromthefield, are in place.By manipulating test environments based on real-time data trends, multiple situations can be calculated. This can be done in simulated environments. This makes it quick and eliminates the use of expensivephysical test environments that requiremaintenanceaswell.