Question

I’m confused about how binding works for statically scoped variables in nested subroutines.

proc A:
      var a, x
       ...
          proc B:
                var x, y
                  ...
                proc B2:
                      var a, b
                      ...
                end B2
          end B
                   proc C:
                     var x, z, w
                         ....
                     end C
end A

First, this is what I have understood: if static scoping is considered, then B2 can use the variable x and y present in its parent B. Similarly C can use the variable a used in proc A.

Now, my questions are: are these bindings made during the compile-time or run-time? Does it make a difference if the variables are statically scoped or dynamically scoped?

Answer 1

Until it comes naturally, I find it easy to draw environment model diagrams. They are also pretty much essential for exams and those esoteric examples that are intended to be confusing. I suggest the famous SICP (http://mitpress.mit.edu/sicp/), but there are obviously more than enough resources on the internet (a quick google brought me to this: http://www.icsi.berkeley.edu/~gelbart/cs61a/EnvDiagrams.pdf).

It depends on the language/implementation when/how bindings are done, however in your example the bindings can be done at compile time. In general, static scoping, as the name suggests allows for a lot of static/compile-time binding. A compiler can look into a function and see all references and resolve them immediately. For example in B2, a reference to y can be resolved immediately to belong to the enclosing scope, i.e. that of B.

As per dynamic vs. static scoping, there is a huge difference. Dynamic, as the name suggests, is much harder to do compile-time bindings with, since the structure of the code does not define the references to the variables. Different paths of execution may yield different bindings. You’ll have to be more specific with the question though.