Atoms and Terms

In Erlang, a term is any piece of data the language can represent: numbers, atoms, tuples, lists, binaries, and more. An atom is a constant whose name is its own value — literals like 'ok', 'error', 'undefined', or 'true' are atoms, written in lowercase starting with a letter, or quoted with single quotes if they contain spaces or start with an uppercase letter. Atoms are used pervasively as tags and status markers throughout idiomatic Erlang code.

A tuple groups a fixed number of terms together using curly braces, such as {ok, 42} or {error, not_found}. The convention of tagging a tuple's first element with an atom like 'ok' or 'error' is so common in Erlang that it has a name — the 'tagged tuple' pattern — and it lets a function's caller pattern-match on the tag to decide whether an operation succeeded before even looking at the rest of the tuple's contents.

A list is a variable-length, ordered collection written with square brackets, like [1, 2, 3] or [ok, error, ok]. Internally, a non-empty list is a pair of a head element and a tail list, conventionally written [Head | Tail], and Erlang strings are actually just lists of integer character codes under the hood, which is why "abc" and [97, 98, 99] print identically in the shell.

Real Erlang programs nest these building blocks: a list of tagged tuples like [{name, "Ada"}, {age, 36}] is the classic Erlang way to represent a simple key-value structure before maps became common, and a tuple can itself contain a list, as in {student, "Ada", [maths, physics, chemistry]}. Because every term has a well-defined comparison order and equality rule, terms can be used directly as map keys, sorted with lists:sort/1, or matched precisely in function clauses.