New Jersey needs your help

https://qz.com/1832988/covid-19-results-in-new-jersey-desperately-needing-cobol-coders/

A greybeard I worked with talked about doing a lot of COBOL work back in the early days of his career (late 70s). He said a little while back he decided to look up some learning resources to brush off the cobwebs since some friends had told him about some lucrative contracting opportunities, but after some review he said to himself "nope, never again".

I took cobol at A&M, and that was enough to know I hated it

My dad applied.

I took COBOL at A&M as well. It was a crappy language that was only used by BANA majors (minored in BANA) . It was used by businesses because it had all sorts of formatting options to insure that long columns of spreadsheet like data would line up nicely

COBOL relics were in high demand right before Y2k as well

91_Aggie said:

---

COBOL relics were in high demand right before Y2k as well

---

Yeah I knew a BANA major who made pretty decent cash in 1999 because he somehow understood COBOL. I'm amazed anything still runs it 20 years later.

The Fife said:

---

91_Aggie said:

---

COBOL relics were in high demand right before Y2k as well

---

Yeah I knew a BANA major who made pretty decent cash in 1999 because he somehow understood COBOL. I'm amazed anything still runs it 20 years later.

---

NASA has things in COBOL as well... We're quickly approaching the point where all those programs are just going to be black boxes. If anything goes wrong, anyone depending on them will be SOL until a new solution can be created.

I think its hard for big bureaucracies have a really hard time thinking of IT resources as things that depreciate and need to be updated and replaced periodically. It's like trying to run a factory that depends on a 100 year old machine. Some day its going to die for real and you'll be at a dead stop for a while.

Edit: I'm having a funny vision of someone using the modern internet to RDP into a terminal that is capable of managing COBOL applications...

I was a BANA major at A&M and took Cobol. My job coming out of A&M was as a Cobol programmer. I was very glad when I transitioned out of that into something else. I used to have nightmares about that stuff.

C@LAg said:

---

91_Aggie said:

---

I took COBOL at A&M as well. It was a crappy language that was only used by BANA majors (minored in BANA) . It was used by businesses because it had all sorts of formatting options to insure that long columns of spreadsheet like data would line up nicely

COBOL relics were in high demand right before Y2k as well

---

George Fowler for the win?

---

BANA major here. Nice mention. I paraphrase him saying...."COBOL will not die...it will be around in some form and still be called COBOL..." I had his advanced COBOL class. It was the genesis of my hatred for COBOL . I went on to get exposed to the wonders of JCL and Microfocus COBOL....i quickly got away from that and went on to Frontend development and some Java. Still remember killing many a tree picking up my program core dumps in the bowels of Blocker....

One time, I developed an inventory control system on AIX using COBOL and an oracle database, with input from 900 mhz wireless spread spectrum RF hand helds.

True story.

SlobberKnock said:

---

C@LAg said:

---

91_Aggie said:

---

I took COBOL at A&M as well. It was a crappy language that was only used by BANA majors (minored in BANA) . It was used by businesses because it had all sorts of formatting options to insure that long columns of spreadsheet like data would line up nicely

COBOL relics were in high demand right before Y2k as well

---

George Fowler for the win?

---

BANA major here. Nice mention. I paraphrase him saying...."COBOL will not die...it will be around in some form and still be called COBOL..." I had his advanced COBOL class. It was the genesis of my hatred for COBOL . I went on to get exposed to the wonders of JCL and Microfocus COBOL....i quickly got away from that and went on to Frontend development and some Java. Still remember killing many a tree picking up my program core dumps in the bowels of Blocker....

---

Loved Dr Fowler. And he wasn't wrong. The language is still being developed and in use widely. Good COBOL programmers command very good money and it will only become more lucrative as the very old guard retires or dies off.

Sine poena nulla lex.

Sine poena nulla lex.

C@LAg said:

---

Bregxit said:

---

C@LAg said:

---

91_Aggie said:

---

I took COBOL at A&M as well. It was a crappy language that was only used by BANA majors (minored in BANA) . It was used by businesses because it had all sorts of formatting options to insure that long columns of spreadsheet like data would line up nicely

COBOL relics were in high demand right before Y2k as well

---

George Fowler for the win?

---

RIP doc

---

aw. when did that happen.

---

2014

https://www.theeagle.com/obituaries/fowler-george/article_ffac932e-216c-11e4-a7cb-001a4bcf887a.html

Sine poena nulla lex.

I heard in passing that some of the current COBOL opportunities are $1000 / hour. If true just wow.

The issue is not the cobol, its the environment. I probably could still write/fix a Cobol application, but my IMS, vsam and other database usage if beyond rusty. And we won't even talk about fixing/updating the required JCL to run things batch.
Current implementations understand web, modern databases and have been used to keep stuff no one understands running. When I was working for Barclays a few years ago they were still running ATM's with CICS, Cobol over ssl http links.
The real issue is with this who wants to go to/deal with NJ government BS. They would be far better served to call IBM and license more mips/cpus.

Bregxit said:

---

SlobberKnock said:

---

C@LAg said:

---

91_Aggie said:

---

I took COBOL at A&M as well. It was a crappy language that was only used by BANA majors (minored in BANA) . It was used by businesses because it had all sorts of formatting options to insure that long columns of spreadsheet like data would line up nicely

COBOL relics were in high demand right before Y2k as well

---

George Fowler for the win?

---

BANA major here. Nice mention. I paraphrase him saying...."COBOL will not die...it will be around in some form and still be called COBOL..." I had his advanced COBOL class. It was the genesis of my hatred for COBOL . I went on to get exposed to the wonders of JCL and Microfocus COBOL....i quickly got away from that and went on to Frontend development and some Java. Still remember killing many a tree picking up my program core dumps in the bowels of Blocker....

---

Loved Dr Fowler. And he wasn't wrong. The language is still being developed and in use widely. Good COBOL programmers command very good money and it will only become more lucrative as the very old guard retires or dies off.

---

Saw he passed 6 years back. RIP. He was beloved. Remember he would meet you in the Chicken for a pitcher. It wasn't him so much as I didn't take to Cobol. It was also my last semester in school and I was sooo done. In my early career have run across some Cobol grey beards that made a career out of it. And yes I am guessing some of those that can still handle fixing soc7's and IMS/JCL could make some decent money fixing COBOL programs...I remember one job that I had to go on call and would get paged in the middle of the night to fix batch jobs....never knew exactly what to do and called the poor primary to help me . Probably didn't help my outlook much...

Fowler was a good teacher and seemed like a genuinely nice person. He is one of the few prof's there who I can easily recall their name.

Now for his TA who I had at the time (probably around 95 or 96), that was a different story. He was a bit of a nervous, excited, brown noser. It did give the class humor though to watch him. I remember being in the lab once with the whole class killing trees by sending our print outs to the printer... the TA got nervously excited because we were wrapping up and printing our stuff so we could turn it in... he started running around the lab yelling "You're in the queue... you're in the queue" like we were about to win the lottery.

UmustBKidding said:

---

The issue is not the cobol, its the environment. I probably could still write/fix a Cobol application, but my IMS, vsam and other database usage if beyond rusty. And we won't even talk about fixing/updating the required JCL to run things batch.
Current implementations understand web, modern databases and have been used to keep stuff no one understands running. When I was working for Barclays a few years ago they were still running ATM's with CICS, Cobol over ssl http links.
The real issue is with this who wants to go to/deal with NJ government BS. They would be far better served to call IBM and license more mips/cpus.

---

Just got around to reading the posted article and a few others. I agree. They don't have a COBOL problem, they have a capacity problem. COBOL programmers are not going to solve that problem. I would put good money they are running on unsupported hardware and OS and cannot get any support from IBM.

We just cut over from our old EC12s to z14s to keep our SAP environment supported. This was our second attempt after the business killed the first project. We warned them they were running out of capacity and they said "We'll just wrote them a big check if so or get our CEO to call IBMs CEO". We laughed and said good luck. Sure enough they started pegging the CPUs out during monthly closes and they freaked. IBM would not even start the discussion. Just a flat NO.

I suspect the same thing is happening in NJ. They didn't expect this growth in usage and now they're ****ed for being cheap.

I think you'd be amazed how many plants till have old machines running in them*.

Heck, I think the military still, sorta, uses typewriters.

That "sunk cost vs benefit" argument can get hard.

*Thus may be changing now that the phantom income issue has been addressed in the tax code.

I'm 100% certain that they're screwed because they were cheap.

What I'm curious about is what a developer does nowadays to try to mitigate hardware/software/language obsolescence. With all the different database options and programming language options out there it seems like you could pick the most popular languages or database products to try to ensure that you'll have easy access to developers and administrators.

Is there anything else that is typically done? Are there other more modern languages or architectures that could be heading for this in the future?

Will governments and financial institutions ever move to newer systems in significant numbers or are they just going to keep paying COBOL developers more and more money?

TMoney2007 said:

---

I'm 100% certain that they're screwed because they were cheap.

What I'm curious about is what a developer does nowadays to try to mitigate hardware/software/language obsolescence. With all the different database options and programming language options out there it seems like you could pick the most popular languages or database products to try to ensure that you'll have easy access to developers and administrators.

Is there anything else that is typically done? Are there other more modern languages or architectures that could be heading for this in the future?

Will governments and financial institutions ever move to newer systems in significant numbers or are they just going to keep paying COBOL developers more and more money?

---

The issue is scalability. Mainframes are all about massive I/O capability and data security. Despite what is written they are not some old decrepit punch card behemoth from the 60s. They are still extensively developed and use bleeding edge technology. COBOL is still evolving as a language as well with new capabilities being added all the time. If done right you can't beat the performance per dollar and reliability when looking at the volume of transactions these places are handling.

New system will likely suffer the same issues, you throw 160x times the load at it and it will fail in spectacular and interesting ways. New systems advantage is that you can hire people that still know how to tune and optimize the systems they run on. Really recent cloud based systems would also likely allow you to throw resources at bottlenecks quickly. But still unprecedented demand, old or new, is problematic. Look at the systems built for "cash for clunkers" and Obama care enrollment. All new technology, the ones that were actually completed, melted under load and the vast majority never worked at all after spending hundreds of millions of dollars. The failures were not only the low bid offshore technology implementation companies but real common name US technology companies.
And even is it the right call to upgrade systems for load of a once in a lifetime event. This is the place that well designed cloud systems excel at, web throughput bottleneck, turn up more, Insufficient database IO ops, click here to augment. The problem is huge amounts of software is implemented without design, much less design to scale to x100s of normal load. This is why AWS exists, amazon was growing so fast it was impossible to keep up adding boxes for the product needs of the week and they decided to implement a strategy to have huge numbers of generic building blocks on standby and able to repurpose any of them instantly to address a pain point.
Mainframe not the problem, COBOL not the problem, not keeping the knowledge accessible and spending the money necessary to keep the technology at least on the supported list is a problem. I see it in government all the time. For instance College Station had a guy that maintained several critical systems in their infrastructure. Spent many thousands on vendor training on him, several years becoming intimately familiar with their deployment quirks, but they never cross trained anyone for even the most basic issues. So along comes another city and offers him a few thousand dollars a year more and they won't even consider matching the offer. Its not like he got a google offer of 100K a year more (and that could completely happen) but now lets let that knowledge escape and spend 10's of thousands to get someone new trained and they will be up to speed in a few years. And in the interim we will hire consultants for multiple 100's $ hr to come in and fix issues, taking far longer since they are not familiar with the deployment scenario.

Bradley.Kohr.II said:

---

I think you'd be amazed how many plants till have old machines running in them*.

Heck, I think the military still, sorta, uses typewriters.

That "sunk cost vs benefit" argument can get hard.

*Thus may be changing now that the phantom income issue has been addressed in the tax code.

---

Spot on. I work in control systems software. I'm never surprised to see very old machines and operating systems in the plants. Many facilities still run systems on Windows NT. Many use various Unix distros that are very old.

I can tell you from first hand knowledge that a formerly very large fortune 100 company was still running systems on OpenVMS/Alpha up until a year ago. VMS was first introduced by Digital Equipment Corporation in the mid 70s.

I get it - I don't work with computers - but, getting a new line up/getting whatever quirks there always seem to be in various sensors, etc worked out, is a major PITA - and takes... maybe a year or two, before it can really be trusted... Lots of poorly made sensors and pressure switches, finding where some idiot didn't fully connect a wire in some component, etc...

So, questionable business decisions and management being blamed on technology? I think businesses creating single points of failure is probably a long tradition. I worked for a company that created an in house program that regularly required new templates to be put into it to support work. They let it get to the point that he was the only person in the country that knew how to make these templates (no documentation, no cross training, nothing.) They took that guy out of IT and stuck him in the R&D group... When things slowed down, R&D laid him off, completely screwing over the production side of the business.

I think that personnel problems are what concern me about critical business infrastructure running on systems that are very expensive to support. If you can virtualize the hardware so you don't have to worry about maintain decades old machines you can mitigate some of those problems.

It seems like the unemployment system could have been built around hardware limitations that existed at the time, whereas now (like you said) with modern methods and cloud computing, you could much more easily design a system that could scale past basically any eventuality. That may have been possible with the software side of COBOL (I don't know), but it doesn't make sense to put the effort into coding a program that can handle 10 times what the hardware is capable of.

My question is whether programming languages have stabilized enough that you wouldn't run into the same problem in the future if you picked a language that is popular today.

I've got nothing to add. Just bookmarking this fascinating conversation.

Admittedly, I know next to nothing about computers.

However, especially in a system to process confidential information, I see the merits of an older language.

Hell, I think most of the tech issues/bugs/privacy concerns come from "silliness " in the code, and no real thought to efficient running... Of course, i think forums would be better if they were just green on black, so...

Maybe I should switch to Linux...

Quote:

---

idiot didn't fully connect a wire in some component

---

Hired the wrong project manager on that project. Speculate on this comment lack of loop checks and signed off installation and commissioning. I can help...well not anymore as I am doing this for someone else. I can be a first rate a-hole on the job, but when I oversee commissioning, it works and is trusted.

Hey Cassius, bet you and I have much in common and similar experiences.