The issue is replicating the requirements of the old system. It's become a critical part of your business intelligence infrastructure, and it's older than most employees. Odds are high the...
The issue is replicating the requirements of the old system. It's become a critical part of your business intelligence infrastructure, and it's older than most employees. Odds are high the documentation on how to 'run' it is good, but not so good on what it actually does. The people who built it are long gone, the code itself is your only resource unless you're lucky enough to hire some of the old folks at a hefty consulting fee, and even then they won't remember everything. I've seen this firsthand at major companies in the past.
So you're saddled with this black box you can't replicate. You have to reverse engineer it, and that's a tough thing to do to a critical production system. It's easier to just keep the old systems running or try to find a way to virtualize the black box.
This problem isn't COBOL itself, it's simply the age of the system, and all COBOL systems are old at this point by definition. Age breeds complexity and brings loss of institutional knowledge in all but the best run companies.
My understanding is that COBOL was one of the first languages that could do all the things necessary to run the books for financial organization, so it was the one that everyone ran off with, and...
My understanding is that COBOL was one of the first languages that could do all the things necessary to run the books for financial organization, so it was the one that everyone ran off with, and being financial organizations, they aren't exactly eager or willing to update to the newest latest best every year. COBOL becomes the standard because it was the standard.
The issue is replicating the requirements of the old system. It's become a critical part of your business intelligence infrastructure, and it's older than most employees. Odds are high the documentation on how to 'run' it is good, but not so good on what it actually does. The people who built it are long gone, the code itself is your only resource unless you're lucky enough to hire some of the old folks at a hefty consulting fee, and even then they won't remember everything. I've seen this firsthand at major companies in the past.
So you're saddled with this black box you can't replicate. You have to reverse engineer it, and that's a tough thing to do to a critical production system. It's easier to just keep the old systems running or try to find a way to virtualize the black box.
This problem isn't COBOL itself, it's simply the age of the system, and all COBOL systems are old at this point by definition. Age breeds complexity and brings loss of institutional knowledge in all but the best run companies.
My understanding is that COBOL was one of the first languages that could do all the things necessary to run the books for financial organization, so it was the one that everyone ran off with, and being financial organizations, they aren't exactly eager or willing to update to the newest latest best every year. COBOL becomes the standard because it was the standard.