It seems like a strange problem to affect a lot of programs (for instance, 
fixed-length in program, variable-length in the JCL).

Assuming that all the OS/VS COBOL programs work, I think it would be reasonable 
to expect that all the Format 1 DSCBs match the definitions in the COBOL 
programs.

To identify where any "DCB information" in the JCL does not match the program, 
it could be "simple" to match the DSN and DCB-info from the JCL to the DCB 
information from the "catalog".

Not so simple, with symbols in DSNs, although depending on your naming 
conventions it may be possible to "normalise" the names for matching without 
relying on symbol-substitution. Accurately stripping DDs from JCL can be fun 
too.

I'd assume mostly it will be DISP=NEW, perhaps some MOD, but note that any 
DISP=SHR/OLD with DCB-info you'll probably want to know what they are up to.

You could then look to tie the DSNs to programs using SMF records.

Assuming your programs are working (as OS/VS programs), I'd assume the fixes 
you are making are to the JCL, rather than to the programs themselves. I'm not 
sure you can do much scanning of the COBOL source to identify the issues. The 
COBOL programs are just writing (or reading) what they need to, it is the 
DCB-info in the JCL that can be "getting in the way".

Is the task large enough for you to need to do someting? How many fixes have 
you needed, compared to how many have progressed without needing fixes, and how 
many programs you have in total? 

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to