What programming languages make you irreplaceable to an employer?

Is COBOL one of them?

Other urls found in this thread:

clojure.org/community/companies
clojure.org/community/success_stories
blog.cognitect.com/blog/2015/6/30/walmart-runs-clojure-at-scale
twitter.com/NSFWRedditImage

Some one answer please. I'm teaching myself cobol. The books old 1989. So I hope I could land a job that's irreplaceable.

Perl, I guess, maybe Delphi.

None, really it's mostly about knowledge of domain and codebase.

Keep it up, we need someone to replace all of the dinosaurs.

Good call, more to add to the to-do list.

Wouldn't something like COBOL at least make you a more attractive asset (if you were skilled within it, of course)?

>Wouldn't something like COBOL at least make you a more attractive asset (if you were skilled within it, of course)?
Nah, I happen to know a 30-something old programmer that learned COBOL first.
He stopped working there/got laid off (I have no idea) and now works for a shitty salary as PM/C#-Programmer.

According to others knowing COBOL first might make you unattractive - as in both sexually and to employers.

COBOL is one of them
I'd say Ada is another if you work in the right department

It seems like COBOL is such a meme nowadays that it's not actually advantageous to know it at all, anymore.

Fun Fact: There's no such thing as being irreplaceable. One, there's like seven billion people on the planet. You think they won't be able to find some other guy who knows (or can learn) COBOL? Two, even ancient legacy shit changes. Who says in five years they don't finally decide to replace their ancient IBM mainframe and dump all the COBOL shit once and for all, leaving you out of a job? Three, it's inherently annoying to employers. Labor is their biggest cost and not being able to get rid of people easily is the biggest thing that increases that cost. They know exactly what you're trying to do, get yourself into a place where you don't have to learn anything or do much work, but are assured of employment. This only increases their incentive to get rid of you as soon as they can.

>According to others knowing COBOL first might make you unattractive
I guess it is better to start off with Python.

>as in both sexually and to employers.
*On your first date in a nice restaurant*
Her: So, user, tell me something interesting about yourself?
You: I'm a COBOL programmer.
Her: Oh, sorry user, I've got to go... Chad is picking me up in 10.
>mfw

What about Lisp?

I was using it figuratively autismo.
Shall I format it so that it doesn't conflict with your pedantism?
>What programming languages make you (WARNING: figurative language coming up) 'irreplaceable' (WARNING: ending of figurative language) to an employer?
Aside from that, thank you for the information, user.

They still use Cobol at my workplace, and they will continue after I die.

>What about Lisp?
From what I gather it's mostly used by academics and hobbyists, and not so much in production.

Not an argument kiddo, go home

>wat is clojure

A lisp-like language that runs on the jvm.

>counting a handful of webshits as professionals

Dang?
>Not an argument.
>Not an argument is not an argument.
Checkmate, Molymemeists.

I was thinking more as a programmer in a research capacity (be that academic, or such)?

What kind of salary do these archmages pull?

Still not an argument. Get out of Sup Forums.

I'm pretty sure he meant legacy stuff like Common Lisp.

assembler

A lisp dialect that is used widely in production

>What kind of salary do these archmages pull?
No idea, but I bet the average is low since it's used in a lot of startups, where devs get paid in equity rather than money.

>What about Lisp?
somewhat, i dont think its as desired as COBOL or Ada though, due to the difference in applications (bank/financial sector for COBOL, military, aeronautics, safety-critical for Ada) but from what I hear its still very much used in academia and research

>Still not an argument.
>Is still not an argument.
Get out of Sup Forums.

Underrated.

Understandable.

Thank you, this is the sort of answer I was looking for.

>A lisp dialect that is used widely in production
now back to reality

you can learn to read and program most languages in a few weeks. knowing a language is always better than not knowing it, but knowing your shit about the systems and problems you're dealing with is more important and harder to replace than "knowing a language"

in-depth expert level knowledge excluded. then again, once the language gets replaced, you become replacable.

All programming jobs will be filled with indians. Become a devops/sysops. Noone in their right mind would let shithands touch their production servers.

Not programming language but I found put that people with proper vhdl and tcl knowledge are valued highly.

clojure.org/community/companies

How do I make the move to devops? I like being a dev right now but I want to transition over the next 5 years or so.

This is very true.

I did think about this actually.
Besides, they'll just end up replacing the Indians with automation.

Thank you for the extra information.

>bunch of memes and companies that obviously only use it for throw-away scripts (e.g. in "labs")
I bet there is a similar meme list for Hasklel.

Fortran
COBOL
ALGOL
Ada

Languages once widely used and still are, but no one learns them anymore.

>What about Lisp?
Every programmer worth its salt learns a Lisp language. Even if they don't know it such as using Clojure. That is you're replaceable.

>haha I know you've got a list of organizations who use Clojure in prod but I'll dismiss it without any evidence

VHDL is mostly in europe though, the americans use mostly Verilog from what I know
not sure about the chinks
good HDL devs are valued quite highly in my industry because you probably will want to kill yourself when "debugging" (i.e. simulating and looking for errors and figure out what and why happened)
good knowledge of Ada and/or C is also a good addition to knowing VHDL

>organizations

not an argument

>I was thinking more as a programmer in a research capacity (be that academic, or such)?
Academic programmers are basically free to use whatever they like, also their code sucks, but it works (good enough).

just like your list of "highly depending users"

>makes claim
>evidence is shown which contradicts that claim
vs
>quoting a single word from a sentence
hmmm yes these are the same

None. The archmages that sit in some dimly lit basement room of corporate HQ pulling astronomical wages as COBOL devs don't get paid because they just know COBOL. They get paid and are irreplaceable because they know the system they're working on like the back of their hands.

Would you rather pay astronomical wages to a few devs that are working on a system that is instrumental to your business and know it's inner workings perfectly or lay them all off and replace them with cheaper college grads that happen to know COBOL but don't have any idea how the system is built and why certain design decisions were made.

>evidence is shown which contradicts that claim
not by a long stretch
Show an actual source that those companies have important software systems depending on Clojure instead of some irrelevant build script/web script memery built by a third party that is also conveniently listed in that list, like e.g. the Walmart contractor.

Because 2 person web memer contractor scripts == will be replaced next year.

I'm also curios how many of those faggets in that list still exist beyond their website.

>"If I learn only this one old tool that nobody uses anymore, surely someone must hire me! :^)"

Here's a cute list
clojure.org/community/success_stories
Here's a post about Wal-Mart using Clojure for some huge system of theirs
blog.cognitect.com/blog/2015/6/30/walmart-runs-clojure-at-scale

So, in other words:
It isn't about having vast amounts of publicly sourceable knowledge (books), but rather, it is about having vast amounts of experience?

No, you cock.
See: in reply to your pedantry.

>It isn't about having vast amounts of publicly sourceable knowledge (books), but rather, it is about having vast amounts of experience?
^^^^^^^^^

So basically Atlassian and Base2 are the only ones that have ever built something relevant in Clojure.

>long ass list of companies
>ONLY THESE TWO ARE RELEVANT
Are you even trying?

>long ass list
>like 20 or so
>most of them are obviously bullshit and/or small webshits and filtered by my brain automatically
>some are "to enable Java team"
Tell me again how any other of those is remotely relevant.

>common lisp
>legacy
Common lisp has support for all sorts of shit, and has very good gui libraries. not legacy quite yet

Which ones exactly were irrelevant? Please explain which ones were irrelevant and why. No memes pls

Ok. Most of that shit - like the Walmart thing - is unimportant web stuff that will be lost like tears in the rain next year. In case of Walmart, I'm pretty sure they already replaced parts of it with the node cancer.
The rest is mostly the same cancer, but made by smaller web companies.

So the only software that will probably exist past 2018 and therefore remotely matter is the flight management software and Atlassians software.

>web
>unimportant
Unemployed person detected
You've got no idea what you're talking about, thanks for bringing down the quality of the board

Scala, Javascript, Python, PHP, C++

>le web is great
>le 8th world wonder
Whatever, the point is, web software has the lowest requirement and the lowest lifecycle.

ITT: old people crying their old languages aren't used and taught anymore

lel

Web isn't great, it's a total mess, but it's definitely important. It's not a pleasing design, which is why so many autists hate it. I'm using it right now to teach the retarded from across the world.

>but it's definitely important.
But that doesn't make backends written for it important.

You're very wrong. The tools people use to make this fucking mess work are important. If they weren't, people wouldn't be working so hard to make a good backend.

>If they weren't, people wouldn't be working so hard to make a good backend.
They don't.
And those tools get replaced by the fad next week anyway, so they don't leave any kind of impression. The web is about the content alone.

Underrated.

>Implying the architecture doesn't matter.
Sure, the current web is a fickle mess and only conceptually important.
However, what would be noble, in so far as much as it would prove the qualities of proper network engineering. Would be to improve the structure and with it the efficiency of the global public network.
>By the way, I'm not the same person you've been arguing with all thread. - I apologise for it accidentally saving in JPEG format and not PNG.