Hello World

?-
write('Hello World!').

=>
Hello World!   true.

Now that we understood how prolog works, let's revisit Hello World.

What is happening here?

Didn't we just learn how sets of Horn clauses are solved by backward chaining? We the solver calculated valid variable bindings to solve our constraints. But where is the output of write coming from? In other languages we maybe wouldn't ask this question since we are used to manipulating global state, memory addresses and registers; there it is normal to call low level libraries and syscalls. But in prolog?

write/1 is a builtin. It calls $write_term, which is implemented in the prolog interpreter itself.

Prolog provides the necessary builtins for I/O, so you can access files or sockets (if you want write a webserver in prolog), …

This I/O-builtins are triggered by the prolog engine when the predicate is evaluated.

In contrast to most programming languages, prolog allows pure declarative programming without manipulating state. This enables very flexible, testable, reusable and even provable code. However, when needed you can include well encapsulated I/O-builtins).