Data Driven

Data-driven tests have several advantages:

• Inputs can be created using other than manual methods. For example, they can come from legacy data, or be generated by models or external oracles. Here is a perl program that generates model constrained randomized data for the CalculatorExample. Note that in this case, the constraining model runs on a completely different computer: http://home.c2.com/~ward/hp35.cgi.

• Tests are easy to review by people (e.g. customers) who may not know how to write the test fixtures or the test generators themselves. They can also create additional tests.

• productivity products like Excel
• scripting languages like Perl, Python and Ruby
• modeling systems like Mathcad or Mathematica.