[open-cobol-list] nested stubs technique ?

[Patrick]

Hi Everyone

This isn't important but I am just a bit excited about something.

Cobol is criticized for not having enough structure. Within a single 
program I thinks it's the most structured language around but perhaps 
with many-many files things could get tricky compared to let's say Ada.

With Ada there are many ways to organize code but one way is the 
package. In the package there is a spec file and one or more body files. 
With the GNAT compiler this could look like this:


airplane.ads
landing_gear.abs
jets.adb
wings.adb

I have some issues with Ada and I am not using it now but it is really 
pleasant to read. You can just look for the .ads extension to find the 
spec file.

The .ads file has a sort of summary of the .adb files. You don't have to 
read the .adb files to have a rough idea of what is going on.

Along this line of thinking I tried this tonight:



       IDENTIFICATION DIVISION.
       PROGRAM-ID. anak.
       COPY "/pat/cobol/nested-stubs/anak-body.cob".


      *Is private to anak scope but initialized each invocation
       IDENTIFICATION DIVISION.
       PROGRAM-ID. anak-word-file-read IS INITIAL.
       COPY "/pat/cobol/nested-stubs/anak-word-file-read.cob".

      *Is public to anak scope
       IDENTIFICATION DIVISION.
       PROGRAM-ID. foo IS COMMON.
       COPY "/pat/cobol/nested-stubs/foo.cob".




       END PROGRAM anak.




The nested programs have their identification division within their 
parent but their code is written in another file. I thought I could 
heavily comment this parent file as well, creating a kind of summary file.

I could imagine using this for many-many nested programs, with many 
levels of scope, much like a Ada spec file.

Has anyone heard of someone doing this before? Surely I didn't invent 
it. Anyone see an issue laying things out this way?

I hope no one minds this not very important email...

Thanks for reading-Patrick