← All docs

Organizing a model: Groups

Another way to organize a model is to place subsets of the model into Groups. Minsky supports grouping, but there are limitations in our current implementation of it: firstly, grouped variables are local to the group; secondly, creating inputs to and outputs from a group creates numbered input and output ports which clutter the model.

Figure 42 shows the strengths of grouping in Minsky :

  • One function can be used more than once in a model. In the top group of the figure, the parameters suit a nonlinear “Phillips Curve”, where the inputs are the employment rate that leads to no change to the wage level, the slope of the function at that point, and the minimum annual change in wages; in the bottom, the inputs are the profit rate that leads to investment just equally profits, the slope of the function at that point, and the minimum level of investment.
  • A plot can be made the interface to the group, and like all plots in Minsky, this is dynamic;
  • It is also possible to shrink a group to save space—see the bottom group in Figure 43—and to magnify it sufficiently that its contents can be seen and even edited directly on the canvas—something that is not possible in any other system dynamics program.

Figure 42: Using the same variables for two different behavioural functions

minsky-quick-intro figure

Figure 43 shows one of the weaknesses: numbered inputs are added to the model as an interface between the top-level canvas and the groups, and these turn up in the browser window. This clutters the model from the designer’s point of view, even though the group reduces clutter for the end user.

Figure 43: Current drawbacks of grouping: extraneous constants created

minsky-quick-intro figure

We hope to release an improved implementation of grouping in Minsky 3.1, which will:

  • Do away with the intermediate numeric variables; and
  • Support global variables within groups, as well as local variables.

In the meantime, we’d advise using groups where they are useful, as in Figure 42, where the same function is used twice via groups, but otherwise use Bookmarks to organise a large model.