Method Category Graph - LucPrestin/Hidden-Modularity GitHub Wiki

The Method Category Graph uses the following mapping between the trace data and the visualization parameters:

The communicating objects are more or less ignored.

We aggregate based on the method sent. Each method has a category that it belongs to. These are what we aggregate on. For each category we create a node with a label. Since we aggregate method categories over all classes that occur, setting the color based on the communicating objects (as we do in the Communicator Graph) makes no sense. Accordingly, this visual parameter is not used. Normalization means that we aggregate certain categories. For instance, there are 20 subcategories of accessing alone, which we combine.

The senders and receivers are used again to create the edges. Here, an edge is drawn from the category of the sending method to the category of the sent message. Multiple edges are merged into one with edge weight. The size of the nodes is calculated using the number of incoming and outgoing edges. The distance of nodes is calculated by the general repulsion of nodes and the attracting force of edges.

Example:

The example is taken from the scenario Opening a Morph in the World.