clojure.spec - Rich Hickey (from LispNYC, 2016)

Testing

"Rich Hickey presents Clojure.spec, a library for defining structural predicates and generating random data to test and ensure robustness in programming, while also improving code maintainability and error handling."

Key takeaways

Spec is a small tool for defining specifications
A spec is a predicate that describes a data structure
Spec is not runtime type checking, rather a way to define and test structural predicates
Structural predicates are a way to specify the structure of a data structure
Generators are used to generate random data that conforms to a spec
Conform returns a new data structure that satisfies a spec
Conform always returns a data structure, not a sequence
One way to look at spec is as a logic library
Spec is not just for single specs, but can be used with collections of specs
Spec does not check for runtime type checking
Spec is about describing what data must be true, not just what is possible
The goal of spec is to be a better way to communicate and work with data, not just to define it
Spec is used to generate random data that conforms to a spec, which can be used for testing
The origin of spec is in logic systems and the desire for a way to communicate about what data means
Spec is used in Clojure to make programs more practical and idiomatic
Spec is not just for data structures, but for functions and arguments as well
A spec for a function is what input it expects and what output it returns
Spec is used to catch typos and errors in code
Spec is used to make programs more robust and easier to maintain
The community wanted a better way to get errors and exceptions, and spec is a response to that
Generators are used to generate random data that conforms to a spec, and testing is a primary use of spec

clojure.spec - Rich Hickey (from LispNYC, 2016)

More talks