COBOL programmers have some of the highest salaries of any other languages specialized programmers, but I don't know if that is due to rarity of COBOL programmers, the fact that those jobs are all government or financial institution employed, or because the average experience for them is 58 years?
The Stackoverflow developer survey debunks this myth, year after year. Not only do COBOL developers make about the middle of the pack, they sit at an average of 18 YOE which makes absolute bank in other stacks
Programming salaries are so inconsistent and these salaries by language become so meaningless.
My buddy who works in Google makes 600k but can also call himself a Typescript developer. I'm a dept lead but I've spent the past few months fixing business critical code, so depending on how the question is asked, id look like a overpriced jQuery/Angular/bash developer.
The idea that a job req could actually ask for "50+ years experience" in a given piece of computing technology just gives me goosebumps. Like someone did a really good job 50 years ago, or a really bad one. Either way, it's astonishing that any one thing could be in production use that long or longer.
When a piece of software does a very limited set of tasks that cannot be meaningfully improved, and when minor mistakes can equate to millions in cash or even lives lost or ruined, the name of the game is maintain, maintain, maintain. It ain't broke, leave it the hell alone, because updating, upgrading or porting your system will inevitably lead to some sort of mistake.
Contractors make a lot of money, but that would be separate to standard engineering salaries
I've known a few people that graduated about a decade ago and decided to work in really niche tech like COBOL, Salesforce/SOQL/SOSL, VB6, Sitecore, etc. Hell, one guy I met was a professional "ActionScript" programmer! Many in-store and company kiosks used Flash to program their interfaces, so he'd do basic maintenance, add features, and collect six figures for half a year of work and all the travel around Europe/Asia he wants.
Most have more. Like 3 guys just learned it as a prank last year for the first time in generations, which kind of threw off the curve. Every other COBOL programmer is technically old enough to retire, but they are contractually required to continue working until the heat death of the universe.
The big users of COBOL are the sort that don't pay their developers all that well. Government, banks, giant corporations. The sort of places that still have pensions for long service
I have some experience and no formal training. If I dove into cobol classes and certs would that alone be enough for potential employers? Not in a get rich quick kind of way, but more of a ‘what’s the fastest way I can become attractive to employers without having to go back for a degree cause my current career is falling apart and I need to transition to something that isn’t actively injuring my body.” Kind of way…
TL;DR: it's probably not that hard to pick up compared to the complex and deep stacks we use today. Someone will give it a shot.
COBOL is in a special place in our computing legacy. It's too new to require intimate knowledge of the electronics that drive it (older systems and machine-code did), and is too old to be all that complicated (target machines were much smaller and slower). I would wager it's actually not that hard to learn, and is probably a dream to code with modern equipment. You won't be slowed down by punchcards, tape drives, time sharing, etc., and can probably use VSCode and an emulator to cover a ton of ground. The computing model is likely a straight line (storage -> compute -> storage), with little to no UI. In other words: simple by today's standards.
COBOL is not hard to learn. But it takes years to develop the muscles in your fingers to the point where you can write a subroutine in a single session.
Fr tho what happens when all the COBOL programmers die off?
Uh, how do you think learning programming languages works exactly?
Maybe we can extract COBOL programmer DNA from a prehistoric mosquito that got stuck in amber, and then combine it with frog DNA to produce a new generation of COBOL programmers? We could build a facility for this on some remote island, maybe call it COBOL Park or something?
People think there's job security in this, but there's really not. I have been called in to replace archaic code with more modern/easier to read code.
It pays very well.
And there have been companies that are paying millions to a small firm to rewrite their COBOL software that covers the same feature set but also opens the door to extendability.
It really depends on management. Some companies don't mind paying IBM for new mainframes just to avoid any risks touching it, others are desperate to "break the monolith" and migrate from COBOL to something modern... like Java8. You win some, you lose some.
Yes, I mean it's used for transactions in the programming sense of the word. Turns out financial transactions require that as well. I assume the same goes for nuclear stuff. There is just very little risk to come across uncertainties when the language is that old (and the people who use it hehe - tbf it pays super well).
Speak for yourself. I don't use LLMs and never will.
It always irks me when people talk about it like it's universal and inevitable when that's very far from the case. There are many, many issues with them and many developers wisely choose to ignore the fad.