Connections

There are two categories of variables in ACT: parameter variables (e.g. pint, pbool, etc.) that are used during expansion/elaboration time to build a circuit, and circuit component/wire variables. While parameter variables are more like normal variables in a programming language with the caveat that in some contexts they are immutable, circuit component variables behave differently.

 one_place_buffer b;

In the ACT above, the statement one_place_buffer b; creates the variable b that is used to refer to all the elements of the circuit component defined by a new one_place_buffer process. Similarly, the fragment:

 one_place_buffer b2;

creates a new circuit component b2 that is also a one-place buffer. So far one can understand this using a conventional programming language view: one_place_buffer is a type that happens to correspond to a circuit component, and b and b2 are variables that refer to those components.

However, where ACT departs from standard programming languages is that you can connect b and b2 as follows:

 b = b2;

At this point, b and b2 refer to the same circuit component. In fact, by connecting the two, the overall circuit only has one circuit component of type one_place_buffer that is referred to by b and by b2. There is no disconnect operation in ACT, so this operation cannot be undone.

Connecting two variables makes them refer to the same circuit element. Connecting two processes makes the two variables correspond to one process (as above), and connecting two channels makes them the same channel. The earlier examples used channel connections to connect the output of one process to the input of another.

The two connected variables can be viewed as aliases, i.e. different ways to refer to the same process. ACT provides flexible syntax to support different ways to connect variables.