If you're unaware of the extent of the world outside your own, it's not because that world does not exist.

COBOL jobs have been a stable source of well-paid employment for decades. The jobs have always been in high demand, pay well, have good benefits and are stable. Working hours are almost always reasonable.

But hey, payroll, insurance benefits, banking, and airline ticketing are not nearly as glamourous as having an unreliable job with poor benefits and long work hours so some kids can be entertained for a couple of hours. Ridicule away and let that feeling keep you warm while searching for your next gig from the wireless in your parents basement where they let you move back.

My amazement is the high demand for Fortran experts these days. A lot of the cool AI is being done in Fortran, and that language predates COBOL by about 15 years. Fortran programming, unlike COBOL programming, does not dull your brain.

Bregma said:

If you're unaware of the extent of the world outside your own, it's not because that world does not exist.

COBOL jobs have been a stable source of well-paid employment for decades. The jobs have always been in high demand, pay well, have good benefits and are stable. Working hours are almost always reasonable.

But hey, payroll, insurance benefits, banking, and airline ticketing are not nearly as glamourous as having an unreliable job with poor benefits and long work hours so some kids can be entertained for a couple of hours. Ridicule away and let that feeling keep you warm while searching for your next gig from the wireless in your parents basement where they let you move back.

My amazement is the high demand for Fortran experts these days. A lot of the cool AI is being done in Fortran, and that language predates COBOL by about 15 years. Fortran programming, unlike COBOL programming, does not dull your brain.

To my surprise, I actually met a young university professor who is writing his physics simulations in Fortran because he can get them to run faster then NumPY or MATLAB. This guy is not a computer scientist, and actually had to put in extra time and effort to learn Fortran.
So yes, there a re a lot of exotic programming niche's in the world

Green_Baron said:

Was trying to be funny, didn't mean to step on somebody's feet, sorry for that. My own world is actually only partly programming (that's just a hobby), much was/is (besides other hobbies like sailing the ocean single handed) geoscience, anthropology, archeology.

This COBOL thing though seems to be mostly a national problem in the US, the skyrocketing unemployment numbers and necessary action to deal with these immediately call for those skills, if the article can be believed. And I fear that many of those who jump the train now may be a record in their own software tomorrow … oops.

Okok, back to things more meaningful :-)

Most COBOL programs should be thought of as queries (Like SQL). While COBOL is a programming language, it is mostly used for old-school data-mining.
As such, like any Query language: Many data-base type technologies are tightly coupled to COBOL. You can compare it SQL which is also slowly dying out to cloud databases. But legacy systems which contain extremely valuable data still use it.
As someone who has had the chance to work on propitiatory data projects: The data (and the system it's on) are usually more valuable then the code that you or I a writing. That's mainely because hundreds of lawyer's legal hours have been spent going over those systems. The Lawyer and standardisation cost of porting such systems is astounding (and usually unknown). Also you would need to pay the legal team per database that you port, so it's not a scalable effort.
So from an economics viewpoint, it's much cheaper to teach a programmer COBOL/Magic/SQL then to port said systems.

Bregma said:
COBOL jobs have been a stable source of well-paid employment for decades. The jobs have always been in high demand, pay well, have good benefits and are stable. Working hours are almost always reasonable.

That said, it is surprising that these systems haven't been ported to newer systems. I can't imagine that COBOL programmers are very common these days, given the prevalence of other languages. I don't work in games, but the main languages I see being used these days are Java, Python, C++/C, and also a smattering of newer languages like Scala, etc. Then again, C++/C and Java are fairly old themselves, so maybe I guess?

Moreover, it is worth noting that a lot of these systems for websites for unemployment claims are buckling under the strain being put on them. There are a lot of stories of systems failing to keep up with demand. Maybe COBOL is not to blame, but it does show that these systems have severe problems…

deltaKshatriya said:

That said, it is surprising that these systems haven't been ported to newer systems. I can't imagine that COBOL programmers are very common these days, given the prevalence of other languages. I don't work in games, but the main languages I see being used these days are Java, Python, C++/C, and also a smattering of newer languages like Scala, etc. Then again, C++/C and Java are fairly old themselves, so maybe I guess?

Moreover, it is worth noting that a lot of these systems for websites for unemployment claims are buckling under the strain being put on them. There are a lot of stories of systems failing to keep up with demand. Maybe COBOL is not to blame, but it does show that these systems have severe problems…

I personally know someone who trains new COBOL programmers.
He mostly takes people who are working in QA and want to get into Dev.
He guarantees them a job at the end of the training.
So everyone comes out on top, and he is in no shortage of applicants. ( and no problem finding jobs for them )
The corporate landscape is very different from the startup landscape.
Also enterprises who use tech (like banks and governments) are very different from even big tech companies.
There are 60 year old legal agreements in existence that are making people boat loads of money that nobody wants to renegotiate. Spending $5m annually on life support for legacy software is nothing compared to what this software is raking in.