Why it matters
HOFs make Scala data processing code compact and clear. Once you know the vocabulary (map, filter, fold, flatMap), transformations that require loops in Java collapse to one line in Scala.
The architecture
A function value is an instance of FunctionN (Function0, Function1, ... up to Function22). Lambda syntax 'x => x * 2' creates a Function1. Method references also work: 'list.map(_.toUpperCase)'.
Common HOFs on collections: map (transform each element), filter (keep matching), flatMap (transform + flatten), fold/reduce (aggregate), foreach (side effects). These form the vocabulary of collection processing.
How it works end to end
Currying: a function of multiple parameters can be transformed into a chain of single-parameter functions. This enables partial application and clean DSL syntax.
Function composition: 'f andThen g' produces a new function applying f then g. 'g compose f' does the same in reverse order.
By-name parameters: parameters marked with '=> T' are evaluated lazily. Used for control structures like 'if' or 'while' that need to defer evaluation.