Effective TypeScript, 2nd Edition

Chapter 3. Type Inference and Control Flow Analysis

For programming languages used in industry, “statically typed” and “explicitly typed” have traditionally been synonymous. C, C++, Java: they all made you write out your types. But academic languages never conflated these two things: languages like ML and Haskell have long had sophisticated type inference systems, and this has begun to work its way into industry languages. C++ has added auto, and Java has added var. Newer languages like Rust and Swift have had type inference from the start.

TypeScript makes extensive use of type inference. Used well, this can dramatically reduce the number of type annotations your code requires to get full type safety. One of the easiest ways to tell a TypeScript beginner from a more experienced developer is by the number of type annotations. An experienced TypeScript developer will use relatively few annotations (but use them to great effect), while a beginner may drown their code in redundant type annotations.

In most languages, a variable has a type and it never changes. TypeScript is a bit different. A variable has a type at a location in your code. The process by which its type changes due to surrounding code is known as control flow analysis.

This chapter teaches you how to think about type inference and control flow analysis, shows you some of the problems that can arise with them, and tells you how to fix them. After reading it, you should have a good understanding of how TypeScript infers ...