In this chapter you will learn
- how to use values to represent side-effectful programs
- how to use data from unsafe sources
- how to safely store data outside your program
- how to indicate your code does have side effects
- how to separate pure and impure code
“… we have to keep it crisp, disentangled, and simple if we refuse to be crushed by the complexities of our own making …”
—Edsger Dijkstra, “The Next Forty Years”
In this chapter we’ll finally address the elephant in the room.
Q I now see how pure functions and trustworthy signatures can help me write better, more maintainable software. But let’s be honest. We will always need to get something from the outside—be it an API call or fetching things from a database. Additionally, almost every application in the world has a state that needs to be persisted somewhere. So isn’t this pure function concept a bit limited?
A We can talk with the outside and still use pure function goodies! The intuition here is exactly the same as with error handling (chapter 6) and modeling requirements as types (chapter 7). We represent everything as values of descriptive types!