The MAIDENS’ implementation distinguishes between generators — an actual software implementation of an algorithm that outputs musical structures — and generator instances — a particular way of using or reusing a generator in the score. The generator nodes that live in the generators container are actually generator instances.
A generator instance is identified by:
1. a binding: which one of the available generators should be used;
2. a target: where the generated content should go;
3. a configuration: how should it sound, with respect to generator’s capabilities.
One can create and use several generator instances for every available generator. While of little use, there is no technical limitation preventing you from having two or more perfectly identical instances — e.g., they all bind to the same generator, are set to work the same way, and route their outputs to the same section of the score.
In the Editor you can set the generator instance’s Binding (choose from list of available music generation algorithms), Output (select one or more of the available section nodes) and Input. If the bound to algorithm accepts an input, here you will be able to select sections or other generator instances — which would chain the algorithms for complex processing.
Still in the Editor, there are runnable actions for running the algorithm (Generate), change its parameters (Configure) or viewing its imprint (Info).
There is a dedicated article on generators with more information.
Note: the complete list of pending bugs, features and epics is available in Jira.