In my continuous journey to design and write better code I try and experiment with multiple testing techniques. A while ago I wrote a whole series of blog posts about Property based testing, an addition to the traditional example-based testing.
A property based test is meant to be fairly succinct and verify that simple properties hold true for all possible inputs given a set of preconditions. Libraries like FSCheck and CSCheck by randomizing the inputs for a particular operation.
But what if we need to test a more complicated scenario where it is not so easy to identify the separate properties of our system? This is where model based testing can help us. Instead of randomizing the input for one operation, we execute an arbitrary combination of operations against our system and compare it to a simplified model.
Let me show you how to do this using CsCheck.
I have created a small Counter with two operations: Increase() and Decrease():
We now define a “simplified” model based on our mental model on how we think the system should work:
Remark: The example I using is quite contrived as the real “complex” system and the “simplified” model are in this case the same.
Now we need to specify a list of operations that can be executed in arbitrary combinations:
The last step is configure CSCheck to execute the operations against both the real system and the model and compare the results.
As you can see the test fails because there is a difference between how the real system and the model handles negative values:
The full example can be found here: wullemsb/modelbasedtesting: A small example demonstrating how to use model based testing with CSCheck (github.com)
For a more realistic example using FsCheck, check out this blog post by Arron Stannard:
https://aaronstannard.com/fscheck-property-testing-csharp-part2/