By: BJ Rollison
Organization: Microsoft Corp.
Title: Test Architect
Input variables are a key component of many tests. Using customer supplied data is important, but it usually only serves to verify nominal values. Constantly reusing static test data in the same test case does not provide significant benefit and has a low probability of revealing new information when reused over and over again. But, random test data is sometimes avoided because it may not be repeatable, and random data may contain incorrect elements causing false negatives in the test result. This talk discusses techniques to generate random test data using parameterized equivalent partitions of the data population for both positive and negative testing. This talk will also demonstrate several free testing tools to generate random test data for use in automated tests, and illustrate the effectiveness of this approach to expose certain classes of defects.
Attendees will learn
- How to decompose input data to identify equivalent sets
- How to use C# Random class members to generate repeatable random test data
- How to use freely available test tools to generate commonly used random test data
- How to use random test data to increase the effectiveness of your automated tests