Re: Rewriting SSA. Is This A Chance For GNU/Linux?

On Wed, 02 Apr 2025 18:36:02 -0400, rbowman <bowman@montana.com> wrote:

On Wed, 2 Apr 2025 22:46:00 +0200, Carlos E.R. wrote:
>

So, to the complexity of handling old code, you add the complexities of
translating to another language.
>
Keep it simple: use today's COBOL. Less translation effort. Fewer
errors.

>
The problem may be finding competent COBOL programmers. That leads me to
another question. Presumably the SSA and other government agencies
currently employ COBOL programmers. What have they been doing the last
fifty years?
>
A more important question might be what have their managers been doing?
There is no question a re-write would be very painful and expensive. For a
government agency there isn't a market force to improve so what would
trigger them doing anything to rock the boat?

There is a lot more to those systems than just the COBOL programs.
The data is in EBCDIC, not ASCII. Sequential files can be fixed or variable length.
For fixed, the length of each record is specified in the file declaration, while for
variable, the length is in the first bytes of each record. There is no character or
combination of characters used to represent the end of a record.
There can be ISAM and VSAM (indexed direct access) files, IMS (heirarchical) databases
and data comunnication with MFS for the 3270 style screen handling, as well as DB2
databases all used by those COBOL programs.
The COBOL programs may have ASM subroutines for the parts that had to be highly
optimized.
The systems are constantly under maintenance due to changes imposed by governements
regulations or due to interfaces with other systems such as pc or cloud usage.
Regards, Dave Hodgins