I'll refute myself a bit.

Our codebase is primarily C++, and we've been looking for people for months. Granted these are more senior positions.

Lots of people out there who can write a fancy resume, but can't design or code.

That's a fact.

I'm told regularly that people come in to interview for senior positions and can't code a simple function to add a couple of numbers.

Oracle and MSSQL are pay for DBs
aws and azure have developed cloud atabases that are tremendously more capable.
I'm using azures nosql CosmosDB now.  It's fantastic.  Except being forced to use JavaScript for ACID transactions via stored procedures

Object Pascal -- Anders Hejlsberg's dress rehearsal for .Net. I always told recruiters back in the early 2000s who were looking for .Net developers to go looking for Delphi developers.

It probably runs the bulk of financial institutions. It's certainly our backbone. About 10 years ago we had the great idea to re-write it into a true real-time system. That failed miserably. I've never touched a line of COBOL but I'm sure 50 years of code has to be horrible. We were having a problem finding qualified replacements to our aging workforce so we started our own training system which seems to be working out well. I know one guy went from the mail room to the cubicle. Several others have made career changes and seem to enjoy being COBOL developers now.

There's no silver bullet. Switching costs are real costs. Lots of big iron AS400 and DB2 systems are still in production and IBM hasn't been destroyed. They're worth more now with a tiny percentage of the market than they were back when they owned it.

ETA: Most new developers I've worked with don't have a clue about indexing or storage architecture. Combination problems are real and in spite of exponential growth in storage and processing power, the demand when things are not properly optimized is logarithmic. I suspect the business model behind cloud services appeals to organizations because IT admin ops are expensive and politically charged. It appeals to vendors because they know that they'll get to charge for processing and storage when poor storage architectures hit critical mass and utilization increases logarithmically.

You have 50 years of simple, procedural code.  To not be able to refactor that into usable libraries of a modern framework is nothing more than gross negligence by your management.

that ... is... quite... a ... statement...

i can assure you it is quite possible 50 years of cobol code AINT necessary simple and procedural. i started at a company where half what i coded was cobol and half was TAL (Tandem A Language.. cant remember what the A stands for.. perhaps application). anyways we had some cobol programmers whose code would give you headaches. a huge mash of gotos and massive confusing code blocks. when i 'maintained' their code, i do what a lot of folks do (its a weakness).. i simply rewrote it from scratch. it took less time (for me anyway) to rewrite an entire program than it did to simply modify it for some change. some programs were 30.000 to 100,000 lines of code..... and i would rewrite the entire thing.

I still occasionally encounter specialized applications that still have COBOL.

LOL this went off the rails for a while.

It's been tried, multiple times.

Many of these businesses will never leave it, because the cost of rewriting it far exceeds the cost of maintaining it.  And most people don't write code that will give the reliability that old code does.

There is often wisdom in the "it's been working the way it is for years, don't touch it unless absolutely necessary" mindset. There are almost always unforeseen/unknown problems that arise when trying to port such a system. These can often far outweigh the costs associated with keeping the old system running. This is especially true if the old system is not well documented and the guys who made it are long gone

This isn't to say that it's *never* a good idea, but it should be approached with extreme caution.

I suspect a lot of the problem is that nobody really knows exactly everything it does, nobody who wrote it is still around (I've heard stories of 80+ year old coders being brought back in and paid astronomical consulting fees though), and as a result they're scared shitless to try to duplicate it.

You really have to understand everything to port a 100k line piece of code, and not just the code.  There are so many edge cases it's going to be a challenge for the people who wrote it originally, let alone someone having to try to learn it while they port it.

Some more stuff: A lot of companies are doing a veteran push because of security clearances. Amazon just hired 30ish software engineer apprentices who have completed a bootcamp. We'll train them up to be devs. Msft and oracle are also trying to hire people to try and get .gov work.

Are you passionate about coding or are you primarily looking for a decent career and coding came up as a thing?

If you can go back in: https://www.airforce.com/careers/detail/computer-systems-programming

If you are looking for a job that won't break your body but aren't necessarily all in on coding: https://www.amazon.jobs/en/landing_pages/mil-apprentice

Right now we only have data center and SDE but we will periodically have other jobs for networking or technical support type jobs.

If you want to do some AWS certs, supposedly the SA one helps people to find jobs but I don't have experience with that personally. https://aws.amazon.com/certification/veterans-reimbursement/

Don't forget to mention those jobs require a TS/SCI with full scope polygraph.

Not my cup of tea.

Good read

https://medium.freecodecamp.org/how-i-went-from-newbie-to-software-engineer-in-9-months-while-working-full-time-460bd8485847

CS50 is a great class

The costs continue to scale up as you support and maintain legacy systems like that.  That code may look like spaghetti but it is still procedural code.  It’s not rocket science to abstract that out and code and test that functionality to the same level of reliability as that existing code, this isn’t 1990.  They have a management problem not a technology problem.

They have to pass the SDE bar so they can transfer to other teams after they get hired on full time if they don't want to do the TS/can't get it. It does means that they won't get the 20k for the TS. I imagine that we'll pay a much higher premium at some point since 20k isn't shit for dev pay and it isn't worth it.

They really don't.  The companies have done the analyses on it and found it's cheaper to teach people COBOL and have them maintain it.  There are a lot of variables that don't work the way you probably think they do.

Yeah that’s obvious, but it’s also a risk to the business to not update their technology.  Eventually that catches up to you.

Very true, but that’s a very good reason to start today and not tomorrow.

It depends entirely what he wants to do. Having a deep knowledge of what is going on at the machine level is absolutely helpful (or even required) in certain jobs. I'd seriously recommend having at least a conversational knowledge of it. Can you be successful without it? Sure, many people are.

I work in the sector, it’s not my first rodeo.

Without knowing their business I can only assume that they are looking at that tradeoff and have decided (for now) that the risks of trying to change are greater than the risks of not changing.  Eventually that balance will shift, but evidently not yet.

I want to make an app for phones...what's the best coding for that?

Every business is always facing competitors with newer technology.  You need to be managing for tomorrow.  And tomorrow always comes quicker than you want.

Which phones, what type of things do you want your app to do?

If you've never written software before you have a steep learning curve ahead of you. Apps are mostly out of my wheelhouse, but Android apps are usually in Java and iOS is using Swift now.

There are schools and Udemy courses that claim to take you from no experience to competent, but I can't vouch for any of them.

Take picture, add description with keyboard, click email button and it send a formatted (simple) formatted PDF.
Interesting...

Are you wanting to do iPhones and Android phones?  Do you own a Mac?

hook me up man - all my clearances have expired but theres nothing i wouldnt be able to get.  Ive been using my CCNP daily since early 2000's, (even passed the CCIE written once) - I have AWS SA-Assoc and Networking Specialty.  I'm not really a coder as my main competency though but I can do it.  I mean I understand and use Git and Maven with my Java (so its not like im some scrub) and I do most my stuff in Java but have been doing a lot of python lately mainly because thats what seems to have the most stuff out there as far as cryptocurrencies go and thats my latest hobby.  Im doing my own Cloud dev projects now with the cloud nosqldb, IAM, kevvault-HSM crypto, VM,s load balancers, etc etc.

I had no idea COBOL was still a thing.  Apparently there's even a GNU COBOL compiler.

https://en.wikipedia.org/wiki/GnuCOBOL

Swift, using Xcode.  You can have a program running in a very short amount of time.

I absolutely agree we don't have a technology problem. But until the people that write the checks are willing to do so and provide the resources, it ain't broke. We manage huge market share and I'm sure there's a huge fear factor over putting that at risk by screwing something up.

fwiw, I've been beating the modernization drum solid for years and I'm not getting anywhere. Not even on a small scale. We're not ready. I don't like it, but I get it. Ymmv.

Sent you an email through the site, send me a PM with your linkedin if you don't get it. We'll chat. I'm an SDE so not a recruiter but I do have some awareness of the various veterans programs. You can also join operationcode.org and meet a bunch of other vets in tech.

I can attest to this firsthand.

I work out if a COBOL mainframe built in the late 80s so 3 decades now of refined or even completely redone codebases with more configuration than you can believe.

This is the third project I’ve been aware of to replace this system. The first two stalled after the consult came up with 1. Ridiculously complicated/expensive new system 2. Cheaper overlay which doesn’t really fix core issues (data interpretation issues Josh spoke to mostly).

This one is actually going places. We’ve got our first 8 or 10 service calls in dev testing and all the contractual and regulatory stuff pretty much squared away. SIT starts 4/1!

I’m super excited about it, I’ve worked in probably a dozen of these mainframes which are still active and been through 4 or 5 of these replacement projects and never gotten to see the end result.:

There are a number of companies offering money and support to some universities if they get a reasonable number of students to take COBOL classes as part of a CS degree.  It's no secret in the finance sector that there might be a sweet modern app, or a bleeding edge website, but a lot of the real work is still getting done with some old, old code, and someone has to keep working on it.

Wondering whats up with the hate for pi's and python in this thread.  Gotta have .22's and 50 BMG's in the safe, you know.  Not every project needs a gear budget, or a project manager, sometimes you just smash a proof of concept together, and decide if it's all worthwhile anyway.  If you always take a "do it right, or don't do it at all" approach, you're just not doing enough stuff, and are getting left behind.

Or you’re doing it right, and that’s critically important in industries not driven by the latest fads.

Python and Javascript and learn AWS.  Get an AWS certification.  https://acloud.guru I learned and passed the Developer exam via acloud.guru.

Udemy has some cheap classes as well.

Ada.

Welcome to the world of IT.  By the time you do figure out it the rules and technology will have changed and you will still be scrambling trying to figure out what is going on.

I am a senior PM that manages developers. Been in this crap for 25+ years. Done it all - programming, analyst, DBA, architect and PM. Worked at Bell Labs and was a VP on wall street at one point. A few questions if the OP doesnt mind answering.

How old are you?

What is your rough geographic zone (northeast, southeast, west coast, etc...)?

The reason I am asking that is both will impact your ability to find a job.

First off - coding is a young man's game and there is blatant age discrimination in the field.

Second, here is the brutal reality of being a developer in the US now - unless you have very specialized skills you are going to be competing with Indians who are here in the US on H1B visas for entry level work. On top of this companies have outsourced tons of jobs directly to India.

Ever wondered where all the Indians have come from that are now in the US? The H1B thats where. They came here on H1B, than got green cards, than brought over their families. The Indians have pretty much taken over the IT field outside of specialized niches like DoD work. In pretty much any company you go into today almost he entire IT department will be Indian. In some geographic zones there will be less of them but here in the northeast and west coast there are tons of them. They are extremely tribal and nepotism is everywhere. They will hire one of their own before they will hire you.

Am I saying its impossible? The answer is nope. But you need to understand what you are getting into before you spend the time and money trying to get into the field. I know tons of young people who have spent the money and than cant find jobs. Than they go end up doing something else pissing that money down the drain.

The language itself is irrelevant. Bright people can code in any language. The most popular by far at this moment in time is Java. I would avoid Java as you will be competing with tons of other developers. If you really want to get into IT, I would do IT security or networks. Everyone needs security and companies are paranoid about it nowadays. They dont like handing their security over to some H1B (but some will do it anyway). If you are in the DC area there is tons and tons of DoD work - I would focus 100% on security certs and than go from there.

I would also learn the fundamentals before worrying about the language - learn operating systems. I mean really learn how Windows, Linux and others really work. It makes it much easer to code because you understand it. I have a CS bachelors and masters. They were helpful because I know how the machine works.

If I was young and had to do it again I would go learn into a licensed trade like a plumber, electrician or HVAC guy. Yes it can be tough on the body at times but sitting on your butt all day introduces a whole other set of problems. Regardless where I live the plumbers make more than surgeons.

Feel free to PM as I will be glad to help you best I can. The IT job market is hopping right now because of Trump cracking down on the H1B, so now might be a good time to get in if thats what you want to do.

I can't think of many industries that don't have some sort of emerging tech in an incubator somewhere.  I work information security in the financial sector, which is pretty stable and dull when you get to the core business, but there are plenty of us working on interesting research items that if good, get spun into full scale projects.  I know enough people working in even more serious companies and .gov organizations that are doing similar stuff, so it must just be  matter of exposure to recent research stuff.

The stuff we all use to protect lives, energy, logistical, and financial infrastructures, or run core business on?  Yeah.  Serious gear, serious code, serious process, and tons of audits.  The "what if?" stuff, you can roll your own wacky crap out of whatever is lying around. It's why we have development labs, test labs, and integration labs.  Some of our "what if's?" fail miserably, but they give us a lot of information we need in order to prove or disprove a concept that sometimes become very large and serious projects, or take a project that was not going to be worthwhile off the table.

No more.

It’s mostly C++ now.

Nobody’s talking about research projects.

H1B don't get security clearances.  The OP already stated he wants to go into defense.

Well, you probably wouldn't be.

I have done that with plenty of them. The discussion was using those devices in production. That’s not research.

They don't ask about you M4M activities any more. You're good to go now.

Guess I will be the nay sayer here. I started in software development in 1980, Mainframe all of this time. I am taking the buyout at my fortune 500 company, after 33 plus years. When I came into my group there were 14 of us, there are 3 now, the positions were back  filled with Indian contractors. At my company it is clear that they have been shifting to contractors for awhile now. There were layoffs a few years ago, lots of Web developers (Java) were let go and I thought Java was the big thing.  I think that system administrators and network communction people will be in demand for some time, coders, maybe not.

If I were a young man starting out now I think I would look hard at Logstics and supply chain. Possibly computer hardware, software, just my perspective, looks like contractors are filling that need.