Technical Presentation
Robust Testing with Probabilistic Stochastic Test Data
Many tests require numerous input variables as test data. Real-world test data is important, but it really only serves to verify nominal input conditions. Also, constantly reusing the same static test data repeatedly in the same test does not provide significant benefit beyond re-verifying those specific constant inputs. But, random data is often shunned because it may not be repeatable and the randomly generated sample may erroneously contain incorrect elements in the data set. This talk discusses techniques to generate random test data using seed values for repeatability and stochastic samples that are probabilistic of the population for both positive and negative testing. This talk will also demonstrate several free testing tools to generate probabilistic stochastic test data and illustrate the effectiveness of the tools in exposing certain classes of defects. Attendees will learn:
- How to decompose variable data into sets to produce random samples that are probabilistic of the total population
- How to use C# Random class members to generate repeatable probabilistic stochastic test data
- How to use probabilistic stochastic test data to increase the effectiveness of your testing
- About freely available test tools to generate valid and invalid probabilistic stochastic test data
- How to identify and troubleshoot defects resulting from stochastic test data
