Each component has a namespace which is distinct from the namespaces in other components.

Component naming is relative - there is no “global” namespace.

For emphasis, I break namespaces into 5 sub-categories:

I choose to emphasize that the input namespace is distinct from the output namespace in all components.

Likewise, /x and /c are distinct namespaces.

Input names and output names can appear to be the same, but do not cause name-clashes (because they are in separate namespaces).

I, arbitrarily, choose the following syntax:

a namespace reference consists of two parts - (1) the component and (2) the namespace within the comonent
a name reference consists of “:” followed by the name.

For example:

comp/i:a

refers to the “a” input port of the component “comp”.

Further examples:

comp/o:a - “a” output port of component “comp”
comp/x:1 - A connection within the component “comp”. We gensym the name of the connection to be “1”. The user doesn’t actually care.
comp/c:sub - The component “sub” in the namespace “c” of component “comp”.
comp/n:xyz - The symbol “xyz” in the namespace “n” of component “comp”.

Optimization and Hiding

Later, we might wish to adopt a UNIX-like naming scheme where all parts of a path use the “/” operator.

For example, we might want to rewrite:

comp/c:sub/i:a

to look like:

comp/c/sub/i/a

but, that is a syntactical bauble that can be applied later¹.

For now, for clarity, I will use the form: comp/c:sub/i:a.

Abbreviating ./

I will use the abbreviation . to mean the current component.

For example ./i:a means comp/i:a when it appears inside of the component “comp”.

This is purely meant for human consumption - writability and readability.

Automation will transform:

root comp
./i:a

into

root comp
comp/i:a