Nestor is splitted into several different libraries (node, base, graph, and retim) that are dependent (base needs node and graph needs base and retim needs graph) and that provide different functionalities. This organization is flexible and allows to use only the needed parts of Nestor. For instance, it is possible to build a syntax checker for HPF codes only by using the node library. If a user is not concerned by program dependences, it does not have to link the graph library. This design approach permits to add functionalities to Nestor in an incremental way without rewriting all the underlying bricks.

In the following, we present in details the different libraries of Nestor: