It’s the 1980s, and legacy systems are already a problem. IBM is still shipping source code for big-iron internals, says this mainframe pilot fish.
“IBM discontinued supporting BTAM under CICS, but we had a major customer who communicated via CICS BTAM,” fish says.
Translation: The big customer uses the CICS transaction-processing system, and connects to it over telecommunication lines using the very old BTAM protocol — and IBM has just stopped supporting that approach.
Boss instructs fish to put BTAM support back in — after all, she does have access to the source code. Fish thinks that’s a bad plan, since changing internal mainframe system code is never a good idea.
But that’s what the boss wants, so fish makes the change and then carefully and heavily documents the change.
Fish eventually leaves the company, but makes sure everyone understands the change and what will have to be done the next time the code is updated.
And the results of fish’s diligence?
“I was told by friends still at the company that after I left they immediately discarded the documentation,” sighs fish. “The first time maintenance was applied, the change disappeared.
“It took a high-priced contractor three months to figure out how to put it back.
“It was one assembler branch instruction. The code was still in CICS — IBM had simply branched around it.”
You can trust Sharky to hold on to your true tales of IT life. Send them to me at sharky@computerworld.com. You can also subscribe to the Daily Shark Newsletter.