Node, port, wire

Describes NodeGX 0.1.0.

NodeGX projects are built as visual graphs, not written as code (though you can drop into code where you need to). Three words carry almost all of that idea.

Node

A node is one unit of behavior or appearance, dragged onto the canvas from the node picker. Some nodes are visual — a Button, a Text, a Group — and appear as elements in your running app. Others are invisible at runtime and do data or logic work — a Variable, an HTTP Request, a Function. Every node has a type (what it is), and an instance (the one you placed) — you can place the same node type many times, each with its own settings and its own wires.

Port

A node exposes ports — named connection points on its edges. Ports come in two directions:

• Input ports — on the left/top of the node — are things the node needs: settings you type in directly, or values fed in from elsewhere in the graph.
• Output ports — on the right/bottom of the node — are things the node produces or announces, that other nodes can react to.

Every port also has one of two kinds — value or signal — which changes what connecting it actually does. That distinction is important enough to have its own page: see Signal vs value.

Wire

A wire is a connection you draw between an output port and an input port. Drawing a wire is how you build behavior without writing code: instead of a Button's click handler calling a function, you draw a wire from the Button's Click output to whatever should happen next. A node can have many wires in and out; the graph you see on the canvas is your program's logic, laid out spatially instead of listed as statements.

Wires only connect compatible port types. Where NodeGX can losslessly turn one type into another at the wire — for example, an object arriving at a text input — it does so automatically so the connection isn't blocked on a technicality; where there's no sensible conversion, the picker won't offer the connection at all.