RE: [Tiny-cobol-users] Re: Dynamic subprogram calls

[James S. H. \(T. C. Users\) <TinyCobolUsers@ZName.com>]

=====================================
But in the IBM mainframe world the default compile option is
frequently NODYNAM.  With that option, CALL "MYPROG" is a static call
and CALL MY-PROGRAM is a dynamic call, since a variable containing the
subprogram name is implied and the name could be changed during
execution.  Interfacing with some products, CICS for example, requires
the NODYNAM option because the CICS API must be statically bound to
the COBOL program.  There may also be performance and maintenance
considerations that determine whether you want to use static or
dynamic calls for a particular COBOL program.
======================================

In the IBM mainframe world it is also important to distinguish between
"compilation" and "linking" or "link editing".

Preparing a program to run is a two step process. The first step,
compilation, translates the COBOL (or any other language) into object code.
But it does NOT resolve issues like the calls.

The second step, the "link edit" step, resolves these linkages.

In the IBM mainframe world there are also run time distinctions between
static and dynamic linkages.

A dynamic linkage references a subroutine that is "outside" of the resulting
code of the linkage process. It must be located at call time, and then
loaded and then executed. In addition, dynamic linkages always being in the
initial state of the subroutine. That is, the Working Storage was always
"initialized" at every call. A common way to force a dynamic linkage in IBM
Mainframe COBOL is to place the subroutine name inside a Working Storage
variable that is initialized with the subroutine name. Even if the variable
is never modified it will not qualify for static linkage because it COULD be
modified.

A static linkage references a subroutine that is "linked" into the load
module. It is present in memory, along with the mainline because it becomes,
thru the link edit process, a part of the load module. Unless otherwise
specified, and particularly in the "old" IBM world, static linkages produced
subroutine that retained their status from call to call. That is, Working
Storage was NOT initialized except on the first call and such variables
retained their values from call to call.


James S. Huggins




...