States need programmers who know old COBOL language to process a surge in unemployement claims

Originally published at:


Bad news, all of those programmers are in the high risk group for COVID-19.


I haven’t written a line of COBOL since I was in junior college - and that was not on a mainframe, so I know fuck-all about things like JCL. Hard pass for me, even if it was 100% remote work.


What is crazy here is the are asking these people to work FOR FREE. Even if it is in high demand right now, apparently there are no state funds for this? It’s not like they could even dangle ‘for exposure’ for a job like this. Also, I suspect the tricky thing here is not even the COBOL, but the archaic rules and logic of the system in general - which I’m sure that institutional knowledge is dwindling as well.


Only in COBOL do you have to write PAGES of code just to add two numbers…


Why do they have so few COBOL programmers?

  • Because everyone keeps saying that they are moving off of it and it’s a dead language. They aren’t, and it isn’t. We just need to accept that there is a lot of COBOL legacy code and that maintenance and continued development will continue in COBOL.

  • Because people who need COBOL programming needs to understand that it’s a less well known language now, and in order to get people to do it, they need to pay nice salaries to do that.

  • Screw anyone who expects coders to work for free. If you want a job done, pay for it. States can procure fast when they need it.

If there is job security and good pay, people will learn COBOL. It’s not rocket surgery to understand this, but they apparently are too dumb to understand that you should never run a system you can’t support.


No shit. Nobody works for free. There is ALWAYS a cost, so to hell with those states asking this. They should have fixed this travesty decades ago.


Any programmer could teach themselves COBOL, and, if they’re good in general,they’ll be good at COBOL in a reasonable amount of time.

Getting people to do that is not just an issue of money, though. It’s also the fact that pretty much your best case job is working for a bank, which is not super appealing, and for a lot of jobs your boss would be someone like the folks in this article. Being people like the folks in this article, they wouldn’t understand why working for them would demand an enormous premium.


PROGRAM-ID. nopenopenope.

working storage section.
01 myrate pic 999 value 500.
01 myinterest pic 9 value 0.
01 net pic 999v99.
Net= myrate * myinterest.


First appeared 1959


The appeal of working for a financial institution depends entirely on your definition of appealing and what you’re chasing.

Sure, you’ll never get start up stock options, overnight millionaire, there’s no super cool mobile app. The pay is probably in the middle, at least if you’re competent. Management is crap shoot, it’s not monolithic across the industry. But, you’re also not likely to run into some “bro culture”. Not likely to be on the cutting edge new technology either.

But, there’s a ton of things you are likely to get around quality of life. From remote work arrangements to very predictable hours, to flexibility within the day. To a relatively stable work environment.

From a technology standpoint, it’ll be super boring. But, when you want to go to the kids soccer game, work a 40 hour week, and have time to spend with the family, it’s all right there.

If you see a financial institution that doesn’t have those quality of life perks and has all the old technology, you’ve found a poor place to work. :slight_smile:


There are a lot of tech companies that have all those things. They’re much more in the traditional engineering and manufacturing sectors than in the hot new SV startup sector. They’re stable, professional environments, you can be at the cutting edge, and your managers are engineers, not bankers.


^ this a thousand times. it takes an afternoon to learn the syntax of a new language - okay, except maybe if you’re dealing with a functional language and are used to imperative one - it’s all of the libraries, the best practices, the subtle gotchas that takes time. and that’s not to mention the program code itself.

im sure there’s probably several hundred thousand lines of code involved. big modern systems ( video games, for instance) can be in the millions.

so let’s hope there’s documentation. even with that, a cobol expert might take months of time to learn and understand a new cobol code base. even if the language itself is not that hard.




Because everyone keeps saying that they are moving off of it and it’s a dead language. They aren’t, and it isn’t.

I was chatting with the keeper of a big govt system written in PL/1. He’s got a job for life, it seems, since apparently all attempts at replacement have foundered.

These old languages are going to be with us longer than the B-52 bomber.

PDP-8’s, Varian V72’s and UNIVAC 1100/2200’s are still out there, still running, still in need of the occasional TLC…


And their beards prevent getting a good face mask seal.


Came for this… leaving satisfied!


A big problem (for IBM) with their mainframe business is that they’re too good. Their customers have machines that have been running perfectly, with no downtime, for 15-20 years, so they don’t see any reason to buy a new one.


Shitty codebase hacked together by generations of developers with mediocre skills, in arcane technology that isn’t suited for the job. Fucked up architecture. Bad documentation, zero automated tests. No git, but often SAP. Strong hierarchy, corporate mindset.

Not a place where somebody who has a passion for modern software development wants to spend their time.


The COBOL programmers I’ve run into over the years were disproportionately represented by cool and tough grandma types who’ve seen it all. Crisis or not, BAMFs like that ain’t gonna work for peanuts. Any government official or bank executive who thinks they can cheap out is in for a rude awakening.