In many software projects quality assurance relies largely on testing. Often, this activity is performed in a crafty way, and assigned to personnel with low qualification. There are even some trendy methodologies that advocate the replacement of specifications by the testing code itself. It comes as no surprise that this prevalence of testing may be perceived as one of the main obstacles to the introduction of formal techniques in mainstream software development.

However, there is a growing interest in reconciling formality in software development with the testing activity, resulting in a variant of lightweight formal methods called property-based testing. In this approach, testing code is automatically generated from specifications rather than crafted by hand.

The talk will deal with the relationship between properties – and, more specifically, types – with testing code. A series of simple examples will be used to present the technology available and, also, some challenging issues in this strange marriage.