Dashboard configuration with #=

Manual contents | Tutorial | Model logic | Platforms | Where NuML runs

The #= operator configures everything the user sees and the clock that drives the run: input controls, graphs and tables, integration settings, and the analytical platforms (which have their own chapter). Directives override the hosting app's factory settings for the components the model mobilizes; untouched components keep their defaults and unreferenced ones stay hidden. Source: directive.js.

1. Directive statement forms

Component #= [directive(args), directive(args), …]   // the normal form
Component #= directive(args)                          // single directive
property #= value                                     // bare runtime property
integration #= [dt(0.25), method("RK4"), maxTime(15), batch, autorun]

Each directive is written param(arg, …) and assigns its arguments to the named configuration parameter of the component on the other side of #=. String arguments may be quoted or, where unambiguous, bare identifiers (legend(Exp) and legend("Exp") are both accepted). Arithmetic in arguments is folded (value(1/3), max(2 * Math.PI)).

Runtime properties can be set individually with the bare form:

dt #= 0.5
maxTime #= 15
method #= Euler
autorun #= true
batch #= true

or grouped in one integration statement, where bare batch, autorun, and rerun tokens act as flags.

2. The DynSS dashboard inventory

A NuML program addresses components by name; the hosting app defines which names exist. The DynSS apps provide:

The four display quadrants are shared: TimeGraph and PhaseGraph both occupy quadrant 00, so use at most one of them (when mixing kinds, the convention is TimeGraph plus PhaseGraph1). Which display receives data is decided by the model's update statements; #= directives only format it.

Other hosting environments expose different inventories — a Project.model program addresses the display and input codes its own Project declares (NuML in Project.model).

3. Sliders and switches

Slider1 #= [legend("Carrying Capacity (K)"), min(50), max(500), value(200), step(10)]
Switch1 #= [legend("Vaccination"), left("Off"), right("On"), value(0)]

Every slider or switch the model binds should carry a legend, range, initial value, and step; unconfigured inputs keep the app's generic defaults.

4. Graphs

TimeGraph #= [legend("S", "I", "R"), title("SIR Model"), ylabel("Population")]
PhaseGraph #= [legend(Exp), title("Iterated Logistic"), color("blue"), mode("markers"), size(2)]

Legend labels do double duty: the analytical platforms name their targets by legend string (sensTarget("I"), stocHist("Harvested")), so give stable, meaningful labels to any stream you may analyze.

5. Tables

Table01 #= [legend("Step", "Population", "Survivors", "Recruits"), title("Dynamics"), precision(0)]

A table shows one row per step; column 1 is always model time, so legend includes a header for it. precision(n, …) sets decimal places per column — if the list is shorter than the columns, the last entry applies to the rest. interval(n) emits a row every n steps. title(text) as for graphs.

6. Cobweb plots

The COB component draws a cobweb (Lemeray) diagram for one-dimensional iterated maps:

COB #= [legend("f(x)", "Cobweb Path"), plot("x => r * x * (1 - x)", 0, 1, 0.01),
        xlabel("x"), start(Slider1)]

7. Integration and run control

integration #= [dt(1), method("Euler"), maxTime(100), batch, autorun]

These override the hosting Project's dt, visitMode/method, and maxTime; the dashboard's method and dt selectors reflect the override. The runtime semantics of dt, RK4 staging, and stopping are specified in the SD manual.

8. Directive reference table

The complete directive vocabulary recognized by the configuration engine:

The platform directives are specified in the platforms chapter. Directives repeated for the same component merge; a directive repeated with new arguments replaces its earlier value.