Stack Overflow Reveals Results From 'Largest Developer Survey Ever Conducted' (stackoverflow.com)

More than 64,000 developers from 213 countries participated in this year's annual survey by Stack Overflow -- the largest number ever -- giving a glimpse into the collective psyche of programmers around the world. An anonymous reader quotes their announcement: A majority of developers -- 56.5% -- said they were underpaid. Developers who work in government and non-profits feel the most underpaid, while those who work in finance feel the most overpaid... While only 13.1% of developers are actively looking for a job, 75.2% of developers are interested in hearing about new job opportunities...

When asked what they valued most when considering a new job, 53.3% of respondents said remote options were a top priority. 65% of developers reported working remotely at least one day a month, and 11.1% say they're full-time remote or almost all the time. Also, the highest job satisfaction ratings came from developers who work remotely full-time.
62.5% of the respondents reported using JavaScript, while 51.2% reported SQL, with 39.7% using Java and 34.1% using C# -- but for the #5 slot, "the use of Python [32.0%] overtook PHP [28.1%] for the first time in five years." Yet as far as which languages developers wanted to continue using, "For the second year in a row, Rust was the most loved programming language... Swift, last year's second most popular language, ranked as fourth. For the second year in a row, Visual Basic (for 2017, Visual Basic 6, specifically) ranked as the most dreaded language; 88.3% of developers currently using Visual Basic said they did not want to continue using it."

Comments

Using Javascript

[phantomfive]

Those poor programmers using Javascript. What a lousy language.

(If anyone wants to know why, I will pick one feature out of many. Say you wrote a large program in Javascript, which is happening more often these days. Then you want to refactor by renaming a variable. In Java or C or C# you can refactor by using an IDE automatically, and if somehow you miss an instance, it will be caught at compile time. In Perl or Objective C or Smalltalk, it will caught at runtime in the worst case. But in Javascript, it might not be caught even at runtime, and instead will just cause strange behavior).

Re:Using Javascript

[BarbaraHudson]

Javascript variables are the way they are so that you can use them on the fly without having to pre-declare each and every variable along with it's type which is a god send not a problem.

If you think that's not a problem, you're the problem. :-)

Re:Using Javascript

[phantomfive]

When you say refactor do you mean variable scope? Javascript variables are the way they are so that you can use them on the fly without having to pre-declare each and every variable along with it's type which is a god send not a problem. You perhaps just need to get a better handle on how the scope works, that isn't really a language problem, that's a logic problem with your mind.

No, when I say refactor, I mean refactor. I can't imagine the state of mind that would even cause such a confusion. The point is, there might not be any errors, because the usage might be entirely legal

Re:Using Javascript

[phantomfive]

That is more of a tooling problem, not so much of a language problem

No, tools cannot be made to fix this problem. For example, say I decide to refactor the variable 'message' to be ' outMessage'. Then in one of my code files, I have this line of code, which I fail to refactor:

obj.message = "I <3 Javascript";

That is perfectly legal, and if I miss a usage of the variable later in the same function, it's impossible to detect without understanding the intent of the code, which no tooling can do. It just gets really tough to write large programs in Javascript.

Of course, you can use Typescript as another commenter pointed out, but then you're not using Javascript anymore, you're using Typescript.

Re: Using Javascript

[cyber-vandal]

LMAO. I shudder to think how awful your code must be.

Re:Using Javascript

[mysidia]

Why don't you just "use strict" for all your code, then?

https://www.w3schools.com/js/j...

Re:Using Javascript

[phantomfive]

It doesn't solve the problem mentioned here, and Javascript programmers do that kind of thing all the time, even in major libraries like D3.

Re:Using Javascript

[Zero__Kelvin]

" I could never afford one of those ridiculous compilers which would create code I could never share properly with the world except those using specific machines with specific operating systems."

If anyone thought, even for a moment, that this person was not serious and not a moron, there is your proof that he is one, the other, or both.

Re:Using Javascript

[phantomfive]

Flow is cool but will not catch the problem mentioned here. If you are using Typescript then you are not using Javascript, you are using Typescript. I don't even understand why that is confusing to you.

Re:Using Javascript

[Zero__Kelvin]

I don't often agree with you, but even though I recognize that YHBT, you are correct. Typing systems are a critical part of any language system, and Javascript fails miserably in that regard. It's like saying in a world where autonomous cars are prevalent "The great thing about this car is that it will allow you to inadvertently drive into a wall at 90 MPH!"

Re:Using Javascript

[Kjella]

obj.message = "I <3 Javascript";

My condolences. Personally I loathe code that messes with variables in far-away objects, if it's a huge program you should call obj.setMessage( "I <3 Javascript" ); I don't know how much time I've wasted trying to track down WTF just did something compared to just setting a breakpoint on or printing a debug line in "setMessage()" to see what's happening from where. Yes, setters and getters are annoying copy-pasta code but it's a wonder for sanity. Same with stored procedures and databases, if something is done from many different places route it through one procedure. Even if it's done wrong, then at least it's done consistently wrong.

Re:Using Javascript

[Zero__Kelvin]

No. When you say "refactor" you mean something else, which wouldn't be such a big issue were it not for the fact that nobody, including yourself, has any idea what you mean when you use the term.

Re:Using Javascript

[phantomfive]

Yeah, actually, requiring setters and getters would be a decent way of working around this particular problem in the language. Good solution. Of course, if you do that, most Javascript programmers will complain and say you are not writing idiomatic Javascript, but that's a problem with the community, not the language.

Re:Using Javascript

[phantomfive]

Plenty of IDEs can already handle this without much of a problem, and as I said, Google closure already does static code analysis which will probably catch any "bad" refactoring you've done.

It's not clear you've understood the problem. The bug is syntactically and semantically correct, so static analysis will not detect it. It's only manifests itself as a behavioral problem. Please try to say something that indicates you understand the issue (if you choose to again reply).

Re:Using Javascript

[phantomfive]

In this case, you can in principle get some of the benefits of both worlds,

There is an alternative principle to consider: it could actually give you the worst of both worlds.

Re:Using Javascript

[lucasnate1]

In Haskell and Erlang you can also use variables on the fly without declaring them, and yet they don't allow one to just refer to a non existent variable and have its value automatically replaced with null or empty or undefined or whatever they call it in js land.

Re:Using Javascript

[BarbaraHudson]

Let me turn that around for you - if it's so important, why did MS wait until c# 4.0? Hint - it's not.

Also, maybe we need a bit more rigour, because too much crap that's released is still obviously in the proptotyping stage. "We'll fix it after release" is bs. Used to be that the cost of distributing bug fixes was entirely on the vendor (duplication and mailing of media) so there was a lot more incentive to get it right the first time.

Re:Using Javascript

[phantomfive]

For example prototyping often goes much more quickly with languages that feature dynamic types

I don't think that's true, but I'm open to data that proves otherwise. Prototyping goes quickly in languages that have good libraries (if I want to prototype a new kind of web server, for example, I might try Golang, because it has good libraries for that kind of thing).

Re:Using Javascript

[Antique Geekmeister]

From painful experience: declaring variables on the fly is wonderful to start out when adding features to small projects. It _does_ come back to bite you, hard, when something conflicts when two developers add the same variable in conflicting fashion.

Re:Using Javascript

[phantomfive]

Somehow, of course. Maybe you missed that; I explained.

Re:Using Javascript

[angel'o'sphere]

JavaScript is just an umbrella term for people using TypeScript and/or CoffeeScript or other stuff that gets compiled down do JavaScript.
As a 'concept' JavaScript is nit that bad, but the type conversation rules and the possibility to use undeclared variables is a nightmare.

Re:Using Javascript

[PCM2]

On the other hand, if declared variable types are so nonessential, why did Microsoft invent TypeScript, complete with proper support for refactoring in Visual Studio?

Re:Using Javascript

[PCM2]

Of course, you can use Typescript as another commenter pointed out, but then you're not using Javascript anymore, you're using Typescript.

Not strictly true, since TypeScript is a superset of JavaScript. Unmodified JavaScript code is still perfectly legal TypeScript.

Similarly, other people have advocated using strict coding discipline as a way of writing better JavaScript. Sure you can do certain things in JavaScript, but you just don't ever do those things, including ignoring entire language features completely. This alone wouldn't solve your refactoring problem, though.

Re:Using Javascript

[Aighearach]

...nobody, including yourself, has any idea what you mean when you use the term.

Nonsense, I understood him perfectly. He said "refactor" and he meant "code thrash."

And he's partially right; if you're engaging in code thrash, and think it is called refactoring, then you'll benefit from a language with more hand-holding.

Re: Using Javascript

[phantomfive]

I'm seriously doubting you've ever written a large program.

Re:Using Javascript

[phantomfive]

Similarly, other people have advocated using strict coding discipline as a way of writing better JavaScript. Sure you can do certain things in JavaScript, but you just don't ever do those things, including ignoring entire language features completely.

That's true but important libraries (like D3, for example) use that sort of idiom standard. So unless you're just going to rewrite all your own libraries (which actually isn't entirely unreasonable), you're going to run into that kind of problem.

Re:Using Javascript

[Karlt1]

Easy solution: Use TypeScript

Re:Using Javascript

[phantomfive]

TypeScript is in the survey, and came in at #9 in the most commonly used languages. It is, however, #3 in most loved languages.

Re:Using Javascript

[140Mandak262Jamuna]

You fail to note one important thing. By the time the code needs refactorization, you would have changed your job three times. So it is not your problem anymore.

The original code you wrote will live on, forcing companies to offer 5$ discounts to anyone willing to install flash so that your old code could continue to run. Other companies will keep using WinXP64 so that they don't have to refactor the old code.

Re:Using Javascript

[drinkypoo]

Found the problem with your otherwise quite amusing science-fiction story. No companies ever used WinXP64.

Re:Using Javascript

[lucasnate1]

In Erlang this is impossible, In Haskell code is supposed to be small enough so that it doesn't happen.

Re:Using Javascript

[packrat0x]

I have never written a web app in BASIC. Is there a good example to look at ?
This survey is mis-labeled. The title should be "largest web developer survey".

Re:Using Javascript

[Antique Geekmeister]

The risk of "have its value automatically replaced with null or empty or undefined" for on-the-fly variables may be addressed in Erlang. Are you saying that Erlang avoids the risk of _overwriting_ desired data with a separate write to the variable with the same name by a different procedure added by a different developer? Unless all functions are local, or all scoping is local, then I find that quite difficult to believe. It's also quite expensive in system resources when function calls occur.

Re:Using Javascript

[lucasnate1]

In erlang all scoping is local and single assignment. There are special APIs to access global variables and they are very limited. If you try access a local that was not assigned yet, you will get a compilation error. If you try to do it with a global, it depends on the function you are using if i remember correctly, either an exception or null.

Re:Using Javascript

[lucasnate1]

Forgot to say this, but the globals in erlang, unlike the locals, are mutable.

Re:Using Javascript

[syntotic]

Javascript is lovable! You have to do raw C++ first to feel the effect of having the DOM at hand and the way you bundle code in javascript. It is funny.

Re:Using Javascript

[phantomfive]

Is Javascript your first scripting language or something?

Everyone is underpaid.

[0100010001010011]

Oh, you hate your job? Why didn't you say so? There's a support group for that. It's called everybody, and they meet at the bar.

Also what is a 'developer'? I have a Mechanical Engineering degree and don't Identify with 90% of the stuff that comes up when Slashdot or most places discuss 'developers'. I write code for stuff like Simulink Embedded Coder, Vector CANape, dSpace boxes, etc.

Am I a developer? An engineer?

Re:Everyone is underpaid.

[BarbaraHudson]

Better watch it. All those stack overflow knobs are going to hate you. The whole thing stinks of selection bias:

About Stack Overflow

Stack Overflow is the largest, most trusted online community for developers to learn, share their knowledge,and level up their careers.

"I'm going to post on stack overflow and level up my career" said no real developer ever.

Re:Everyone is underpaid.

[Dutch Gun]

Yeah, I've never talked about it in those terms, but we all know bullshit marketing-speak when we see it. On the other hand, I've often said "thank God for Stack Overflow" after finding a quick and informative answer to a technical question I had.

It's an incredibly valuable resource. I often find it useful when I'm first digging into a new language or technology. Nearly every basic or even advanced question I tend to ask has been asked and answered already, and I can just reap the benefits.

But the *real* payoff, in my opinion, is when you find answers to incredibly obscure issues for which you might have to work days or even *weeks* to figure out, and some kind soul who has already gone through that pain shares knowledge for the good of everyone else, even though doing so is even more work for them.

Re:Everyone is underpaid.

[mysidia]

I would be willing to post solutions/answers on StackOverflow,
If StackOverflow would generate revenue for me based on people viewing or using my solution.

Re:Everyone is underpaid.

[Anonymous Coward]

Of course this then raises the question of what's in it for the people who are actually answering the questions?

I can't speak for anyone but me, but I answer questions on Stack Overflow because I find that effectively communicating how something works requires organizing your thoughts in a way that is of benefit to my own work later, and as I get more seniority also helpful in my ability to explain concepts and ideas to others in my workplace.

In that regard, contributing to Stack Overflow does help me "level up", since it gives me somewhere to practice a skill that is critical to being a senior engineer and to cement my own understanding of concepts.

Re:Everyone is underpaid.

[Zero__Kelvin]

It sounds like you are an engineer who develops software .. ergo you are an engineer and software developer. What you ARE NOT, from what you have posted, is a Software Engineer. I say this because you said "I write code for ..." not "I design software for."

I am a coder. I am a software developer. I am a Software Engineer. What is the difference? For any system that involves a non-trivial amount of complexity or is not lacking importance competent people design before they code. They document their design. 90% of system development is design. Software is the only industry I know where someone says "I need a car that performs like this" and 90% of the people in the industry say "Sure. I have a a bunch of wrenches and other tools right here. I'll start building it immediately!"

Re:Everyone is underpaid.

[Zero__Kelvin]

If you had a solid history on Stack Overflow it would certainly help to be able to point to it. I'm a developer, and I say so, thereby making your "said no developer ever" claim false. Sorry I have to disagree with you. Having agreed to you twice today I was hoping for a trifecta!

Re:Everyone is underpaid.

[Zero__Kelvin]

"If StackOverflow would generate revenue for me based on people viewing or using my solution."

I think we would all be OK with that as long as we could all get paid when you offer a "solution" that is stupid and/or wrong. I know I could use the significant income such an agreement term would generate!

Re:Everyone is underpaid.

[BarbaraHudson]

I don't think HR even knows what stackoverflow (and certainly not what a stack overflow) is. But seriously, have you really said that you were "going to post on stackoverflow and level up my career" to anyone? It sounds so much like what kids who want to break into coding by being game testers would say.

Re:Everyone is underpaid.

[0100010001010011]

For any system that involves a non-trivial amount of complexity or is not lacking importance competent people design before they code. They document their design. 90% of system development is design.

Well, that's because I wouldn't say I "write" Simulink models, but vernacular aside that's exactly how most modeling work is done.

Now we get to get into the 'no true scottsmen' question of if Simulink modeling is coding.

Re:Everyone is underpaid.

[angel'o'sphere]

You are probably both, engineer as academic title and job title, developer as job title. What is your problem?

Re:Everyone is underpaid.

[angel'o'sphere]

I get job offers via my Stackoverflow profile. So I'm quite certain HR does know what stackoverflow.com is.
However it is quite astonishing that the job offers usually have nothing to do with my expertise on SO, last job offer was regarding GWT, I was hailed as an expert. I doubt I answered more than 3 GWT related questions.

Re:Everyone is underpaid.

[BarbaraHudson]

So basically recruiters are just using stackoverflow as yet another source of names to pester, without any actual understanding. Like linkdin.

Re:Everyone is underpaid.

[angel'o'sphere]

I guess they even pay for it ;)

Re: Everyone is underpaid.

[0100010001010011]

we can all agree Simulink is not a programming language.

Why? Based on the GP we go through at much, if not more, design before we make our control systems.

Every paradigm of programming that goes into written code goes into Simulink modeling.

Re:Everyone is underpaid.

[syntotic]

(*Sigh*) 213 countries? We did not finish introducing Horse in whole areas of the world but we gave Computing for Free to just about anyone. That is a huge fault in the market! Distorting. I do not think looking for natural advantages was so valuable as to distort the market at the expense of innovation and... development. We already had Mathematics to that effect, any talent could have surfaced as mathematician rather than wrap us in just-about-anyone-can-program. Which would have meant real quality and much higher wages and less online clutter, etc. But anyway, our New Theory of the World is to give utter freedom to about anyone then just fend off and see what happens next! Surely those guys are using best methods to crib their survey data and give us discounted ubiased quantities and interpretations, right?

Popular vs. loved

[tomhath]

Rust was the most loved programming language

And yet the use of Rust is down in the noise, didn't even make the chart. Apparently the one guy who's using it really likes it.

Re:Popular vs. loved

[thegarbz]

The two points may not be related.

One can love a programming language while not being able to use it for it's current task. e.g. I really like Java* but I need to program assembly on this obscure microcontroller.

*I don't.

Go Forth and multiply

[Hognoxious]

Swift, last year's second most popular language, ranked as fourth.

What about Forth? I hope it was fifty-fifth.

Re:Take the survey with a gain of salt

[BarbaraHudson]

Yeah, statically typed languages look overwhelming/verbose and take more time to type and plan out

That nasty "planning" - there's no time for it in today's culture, where everything is just thrown together after a few "planning sessions" that are basically verbal diarrhea pushed by "big vision" marketing and bosses who may have had a clue in the past, but don't any more and are flailing about to find some project to justify their jobs, same as almost everyone else chasing the big-money exit strategy dream instead of doing the hard stuff like, you know, planning.

The whole "vision thing" has turned software into the cesspool it is today.

Real developer survey

[SuperKendall]

I'll believe the results of a developer survey when all they release is the specs for a web API to vote, and you have to manually do the REST call to submit.

Re:Real developer survey

[somenickname]

That's too low a bar to entry. It needs to be a socket that hangs off an IPv6 address that specifies its endianness at connection time and requires your connection to handle obscure return values from socket commands. The server may also, depending on its mood, request the client to speak EBDIC or request that it negotiates microsecond time synchronization before accepting any answers.

Re:Copy-and-paste programmers

[BarbaraHudson]

Don't forget, most of them are on stackoverflow because they want to, as stackoverflow says in their PR puff piece - "level up their careers." Feel free to punch anyone in the face who says that. It's not like you'll cause any real damage.

Most compilers are free

[Anonymous Coward]

Most compilers are free if you don't use Microsoft stuff.
And even MSFT gives away a version of their compiler - think it included application tracking, but that is to be expected from them these days.

I haven't needed to pay for development tools since around 1997. IDEs are for noobs. My entire OS is an IDE.

Perl, python, ruby, C, C++, FORTRAN, Ada, Rust, Pascal, Go, Erlang, lisp are all free. Out of 500+ different languages, only about 30 have paid compilers that people actually use.

OTOH, you couldn't pay me enough to write code in javascript. Well, that isn't true, but nobody would pay that much.

Javascript is the new php. Used by a bunch of noobs who have to re-re-re-learn all the things the 'old guys' learned in the 90s. But most of them won't because they are just happy to have something working and stop there. Screw security. That isn't important.

Being new isn't bad. Being new and arrogant is. It takes decades to earn my level of arrogance and I HAVE EARNED it.

Re:Most compilers are free

[Captain Ramage]

You do know that the C# compiler comes with the OS, right? Free. The IDE (Visual Studio) is free (Community). You can get the whole Mono thing going, too. So, you were saying?

Or from the classic TDWTF

[Ecuador]

Or, from the classic TDWTF: MUMPS.

No, Visual Basic has nothing on such "brilliant" languages, in fact it is much more pleasant than many other languages as well - e.g. COBOL.

Re:Or from the classic TDWTF

[angel'o'sphere]

If I had to chose between VB and COBOL, my choice most certainly would not be VB.
COBOL might be verbose, nut it is a fine programming language/paradigm.

I guess you never used it and have not much clue what magic you can do with PIC().

Re:Take the survey with a gain of salt

[Zero__Kelvin]

I am agreeing with you AGAIN! There is no way to create software that involves "post-facto planning". Agile could have a place within the software development ecosystem, but you need a vision with an orthogonal view and the time to present/documebnt it coherently, prior to implementation effort, if you want to produce quality software.

Re:double counting?

[phantomfive]

The survey included dependent territories, so presumably American Samoa and Puerto Rico.

213 countires...

[Anonymous Coward]

Funny that the United Nations recognises 193 countries but this survey includes 213.
http://www.infoplease.com/ipa/A0001295.html

Re:Take the survey with a gain of salt

[BarbaraHudson]

What can I say? Forced early retirement must have mellowed me a bit :-)

Re:Take the survey with a gain of salt

[lgw]

Geez, would you guys just get a room!

Re:Take the survey with a gain of salt

[Zero__Kelvin]

I'm really big on getting to know a person before I sleep with them. I know a lot of people think it is weird, but I like to know, for example, if they really exist and if they are a male or female. YMMV of course ;-()

Closed

[Anonymous Coward]

This issues has been closed because somebody asked the same question.

Think of the maintainers...

[srichard25]

The people building massive systems in Javascript may not mind it too much, but I sure hope I never end up having to maintain one of these massive Javascript systems. We all know it's difficult to come back later and extend a system, even if you were the one who originally built it. Javascript makes that problem 10x worse. Refactoring Javascript consists of running Find across the entire source and slowly going through the long list of results to see what needs to be changed. Hopefully all the developers used consistent names for the same variable. And finding all the "objects" of a particular type? Good luck with that.

Javascript might be fast to initially build stuff in, but it's an absolute nightmare to extend and maintain over time. 3 years from now most will consider this type of system to be an anti-pattern.

Re:Take the survey with a gain of salt

[lgw]

Well, if you discover the answer to that for Tom "Barbara" Hudson, let us know. :)

To summarize

[somenickname]

To summarize, a bunch of dime-a-dozen web guys, who rely on stack overflow for every other line of code, have declared that they are underpaid. And, they would prefer to work at home so that if someone asks them a hard question, they can ask it on stack overflow before answering.

Re:To summarize

[ruir]

Because guys that do work do not post to stackoverflow and live in an alternate reality world, right?
It is because of douchebags like you that slashdot has gone to the gutters.

Re:To summarize

[somenickname]

No one disputes that stack overflow is an invaluable resource for a modern programmer. But, it's also a crutch. And the programmers that lean on that crutch the most seem to be the most represented in this survey.

Re:To summarize

[Anonymous Coward]

That sounds like bullshit. How do you know?

Re:To summarize

[MooseMiester]

I call it "programming by Google". To see what this looks like in the real world examine just about any WordPress website you find.

Re:To summarize

[Anonymous Coward]

No one disputes that stack overflow is an invaluable resource for a modern programmer. But, it's also a crutch.

So any kind of documentation is a crutch? Does that extend to header files?
What about intellisense? Is that a crutch too?
Geez, what about compilers? Any programmer that doesn't translate his own code into bits is just flat out lazy!

Re: Take the survey with a gain of salt

[nasch]

That just sounds like bad teaching. It's not as though packages and classes in Java are hard to explain.