1. What kind of language it is

▶ Open in Aratu to hear and change the examples

Aratu is a declarative language: the text describes what sounds, not a list of orders carried out one after another. It belongs with HTML, CSS or a spreadsheet. You do not write "play this note, wait, play the next"; you write what the music is, and Aratu works out when each sound happens.

Some consequences:

The order of lines hardly matters. A motif may be used before it is defined, a pattern played before it is written. Aratu reads the whole text before playing (in programming, this is called hoisting).
There are no changing variables, loops or ifs. Repetition is already in time: every line goes back to its start by itself. Choice is {c|g}. Conditions are about time: at bar 9:, every 4 bars:.
The text is the single source of truth. Each edit makes Aratu read the whole text again. What the player keeps from one version to the next is little: the bar where each line written live started sounding.

Try it▶ Hear it in Aratu

tempo: 100
melody = | riff riff+1 riff+2 riff |
motif riff = [e4 d c d]

Here the motif is written after the line that uses it, and it works.