Stack Overflow User Survey Results | Hacker News Reader

Stack Overflow User Survey Results

blog.stackoverflow.com

91 pointsjncraton53 comments

I participated in the survey, and I don't expect anyone from SO to actually read this posting on HN, but I did find that it was extremely biased towards the idea of people who work places actually ship software based products/services, and it was also extremely biased towards people that work on the web (see the "what are you most excited about" category, which was mostly web-based technologies).

I'm a software developer by day, and I work with a team of around 20 or so other developers. But we don't ship products or create for the web. We're a robotics research laboratory, rooted pretty squarely in academia and non-profit. The closest we ever get to a "product" is a public YouTube video demo or a publication and a conference talk.

I do a little bit of web/app stuff on the side, for fun, but I took the survey from the perspective of my professional programming life, which is my day job.

I know that people like me are in the minority, and that web dev/app dev is where the growth is right now, but it seems a little silly to skew a survey like this away from "people who write software in the general case" because then you aren't really creating a survey of all StackOverflow users. I actually found it a little challenging to answer some of the questions accurately. Seeing how they tied their results so strongly in to their Careers product makes the layout of the survey self-evident in hindsight, but I was a little unsatisfied with the survey itself regardless.

What language do you use stack overflow for?

I use SO in a lot of different ways. I enjoy looking for questions that I can answer in any of the languages that I have experience with. I also use it as a simple search engine for inspiration; if I have an implementation task I'll often search SO first with some of my ideas to make sure I'm not reimplementing the wheel. All of our development at the lab is in Java, which I've been using for about 4 or 5 years on-and-off, so I don't often find myself asking a lot of "how do I do this?" syntax or technicality questions when it comes to the day job stuff; it's usually a cursory search for details or an algorithmic type question.

As far as actually asking questions, I've been on SO for a few years so I've used it for a lot of different things over time. I used it when I was first getting acclimated with C and Java outside of class projects, and now I use it whenever I'm picking up a new pet language. I've been asking and searching a lot about Haskell lately.

That's interesting, I teand to get to SO as a result of a Google search rather than searching directly.

Also I find most questions are simple implementation issues or bugs, rather than around complex design problems. Perhaps I am missing a trick...

>Just like the printer post above, it's incredible how many webcams are left completely unsecured.

Likewise. I started noticing that SO results to software-related queries were of sufficiently better quality that I started to qualify some searches with site:stackoverflow.com. (Similarly, I also add site:ycombinator.com with some other types of searches.)

It's too bad that SO/SX has fragmented their URL structure to make the site: qualifier less effective. E.g., it would have been better to have developer.stackexchange.com, superuser.stackexchange.com, etc. Then you could choose to do just site:stackexchange.com to search the whole corpus vs site:developer.stackexchange.com to be more narrow.

I think it's pretty accurate to say that most questions are, since SO tends to lean towards people who are still learning.

But the high-level questions are out there. Just gotta dig.

The info-graphic cheekily dissolves into an ad, which I guess is OK given the context.

I'm surprised to see JavaScript as the most popular language, but pretty pleased.

Living in New Hampshire I don't have many programmer friends, so communities like StackOverflow are important to me. Relatively new additions like the SO JavaScript chatroom[1] help too.

It may not feel like a community in the traditional sense to some, but in a lot of ways StackOverflow gives me the encouragement I need to help people, practice my craft and keep learning.[2] It really means a lot to me (and I'm sure a lot of others).

Also, note that the tiny link below the info-graphic has the raw data if you want to take a look at all the numbers[3].

[1] http://chat.stackoverflow.com/rooms/17/javascript

[2] http://i.imgur.com/POZmt.png

[3] https://www.surveymonkey.com/sr.aspx?sm=vU4rF_2bPVQaftSo1s69...

JS is #1, and JQuery is #3. That's a runaway win by the JavaScript stack.

Note that the "win" is less about language-vs-language and more about environment. Fifteen years ago, you saw Java score a similar rapid "win" over C++, which is gave up more slowly to C# over the following decade.

What this really shows is that (1) most programming is "IT" programming (i.e. focused development for specific businesses), (2) most IT programming is user-focused, and (3) we've seen a rapid shift in the customer's preferred UI environment.

Java was a revelation for people developing client software in the 90's. It was faster and easier, and in many cases could be deployed over the web. Its later wins on the server backend (which came more slowly) were driven by the client side.

And Javascript is taking that mantle, driven by the mobile world. Yes, you could still develop a Windows client app, but it won't work on your employees iPhones and Galaxy S3's. So you deploy a web app instead.

And thus you use Javascript.

According to the TIOBE index[1], it only shows that SO is very biased toward javascript, not that it is more or less used. StackOverflow is simply the new goto place for js programmers and seems to become more and more exclusively driven by web development, which might not reflect the whole programming industry.

[1] http://www.tiobe.com/index.php/content/paperinfo/tpci/index....

True to a degree, but it's not as if TIOBE is unbiased.

I think it's a bit more complicated than that. Business / or "IT" as you call often has DB + server-side + client-side. The server-side has been taken over mostly everywhere by either Java or its younger cousin C# and you go to any big company and you'll find one of these two languages heavily used.

What is changing now is that people / employees have realized that it's convenient to be able to use any browser, from any device (desktop / laptop / tablet / smartphone), from any OS (iOS, Linux / Android, OS X, Windows) on the client-side.

And the one indisputed king inside the browser to code the client-side is JavaScript.

IMHO It's not exactly as if monothreaded JavaScript engines are going to displace server-side Java / C# anytime soon.

Webapps are accessed from clients using JavaScript. But Webapps themselves, at least for many "enterprisey" stuff are very often coded in Java or C# (or at least virtually always in languages targetting either the JVM or .Net).

I don't think we actually disagree. Backend language choices have always shown more diversity, and will continue to do so. But still, the number of lines of code and developer hours dedicated to these applications is dominated by the front end. With very few exceptions, UI apps have always been "heavier" than the data engines behind them.

So even though C# won't be going away any time soon, it still shows well below Javascript in terms of "acceptance".

Yeah, Javascript must be really confusing to have so many people asking questions about it :-).

I jest, but seriously, popularity of a tag on StackOverflow should be used to determine one thing: popularity of a tag on StackOverflow. There are more important factors influencing a subject's popularity on SO than its overall popularity. A big factor is where the community around the technology tends to congregate. Many projects (like Guava: http://code.google.com/p/guava-libraries/) officially endorse StackOverflow as the place to ask questions, because the maintainers and experts will be monitoring SO. That will have an inflating impact on its relative popularity.

Indeed, mega-monsters like C++ and Python are old enough that Google sends people away from SO to more traditional old books and blog posts which already have most questions answered.

Of all places I'm surprised SO would consider jQuery a language separate from Javascript, an aggregate with a footnote would seem more appropriate.

Not a big deal it just seems odd to me. At least they didn't put HTML up there :)

I was surprised by that too. Maybe they allowed free-form answers.

No, it was a checkbox array (I took the survey). The thing is, saying you use jQuery is not the same as saying you use Javascript. The only Javascript I really use is jQuery. I have no deep JS needs, so I only use jQuery to make interfaces pretty. I chose jQuery (and C#, and some others) but did not choose JS.

Still - you are yet beholden to the quirks of JS like how operator precedence and casting works.

You should have also added JS even though you use the jQuery library extensively.

To be fair, positions 1, 2 and 3 (JavaScript, SQL, jQuery) are "common" languages in that most web developers use/need them in addition to their core server language/framework (apart from node.js of course).

These numbers are not mutually exclusive. The caption for that graph is "% of respondents who have worked with a particular language".

Some high percent of developers in a web company have most likely worked with Javascript. Let's say it's 70%, then that's (roughly) 70% of developers who work with Ruby on Rails, Django, PHP, and ASP.net, among others. This is also true of SQL, which is the number 2 result.

Interesting to see that "Other Europe" is so high with 16.7% of respondents, especially since they have the biggest European countries listed out separately. My guess is that this shows the large number of developers in Eastern Europe.

"Other Europe" is actually just slightly under-represented, not overly popular. All the small countries are pretty big when put together.

Europe has a population of 595M, UK+France+Germany+Italy is 269M, or 45% of Europe. On StackOverflow, the big countries account for 54% of all European respondents.

The "Where do developers live?" chart wasn't particularly satisfying[1], so I thought it would be interesting to look at developer population in proportion to the total population reported by the 2010 census[2].

After doing so, Washington and Massachusetts come out on top. California is #6.

Full details in this google spreadsheet: https://docs.google.com/spreadsheet/ccc?key=0Ak1kLptmHteHdEV...

Briefly:

Washington

  2.15% of total population, but 5.3% of developer population. Ratio of 2.47

Massachusetts

  2.04% of total population, but 4.7% of developer population.  Ratio of 2.3

...

California

  11.91% of total population, but 14.9% of developer population. Ratio of 1.25

...

I threw the spreadsheet together somewhat hastily, so apologies in advance if you catch a typo.

[1]: http://xkcd.com/1138/ [2]: http://en.wikipedia.org/wiki/List_of_U.S._states_and_territo...

> Most popular programming languages: jquery

Yeah, that's what StackOverflow has become.

Yeah that was weird but you can argue it has become a kind of DSL for DOM manipulation.

Yes I know it's not a language per se but because many questions request answer specifically in jQuery and not vanilla JavaScript, 84% JavaScript questions have jQuery tag and only 39% of jQuery questions have JavaScript tag, I think it's pretty fair to include it as a language of its own.

I don't think they cared about pedantic at all.

One metric I would have liked to see was whether people were more or less able to find _good_ answers to questions. A lot of times my questions are answered with "you shouldn't do that" and usually, I know that it is bad, but turn to SO for tricky stuff. I can't change things like the build process or build scripts easily at work for example, I want to know how to adapt to it instead of getting told me how wrong I am.

I mostly code in C/C++ and the number of meaningful answers to my questions has declined over the years. Maybe I don't ask stuff as trivial as I used to but I just can't accept as many answers as I used to, despite community's pressure to "work on my acceptance rate". I end up accepting the answer that helped me the most, even if it was very little help.

Why is jquery counted as a language? Javascript is a language, jquery is a framework.

haha in-joke.

The "satisfaction %" heart graphic made me chuckle. The 1.4% growth from 2010 to 2011 doesn't seem like a big change in heart size, but the 0.8% between 2011 and 2012 - that deserves a huge heart.

Even when produced by trustworthy sources like SO, statistics can mess with your brain.

"You like us…you really like us!"

The problem with this statement is that the survey is based only on SO users. Non-users cannot comment whether or not they like the service.

Sincerely,

A StackOverflow non-user

Do you make a concious decision to not use StackOverflow for a particular reason or is it simply not a tool you need?

StackOverflow is simply a tool that I do not need.

It is not a community site, it is a Q&A site. Too formal. I find it far better to ask a question on a specific mailing list and not on SO. It causes fragmentation and I prefer to have one canonical source of information. Rails question? Ask the mailing list. A community site for developers would be welcome.

Another big problem that I have with SO is moderation. I do not want anyone editing what I have wrote. It would be fine with the occasional editing of egregious errors, but SO moderation is out of control (IMHO).

Someone editing your post and moderation are two separate things.

What aspect of the moderation is gone wild? Noting a difference between those with diamonds (moderators) next to their names and those with a high board score (high rep users).

I would like to see a survey about people's being (un)happy with the current moderation regime, as I have a few bones to pick with it and I don't think I am alone:

1) Some very useful questions such as "what framework would you recommend for parsing CSV in .NET" get closed as "subjective, possibly leads to arguments", whereas they already have a dozen of very useful answers. Had the moderators got there on time, I would not get the info I wanted.

2) Short answers get auto-converted into "comments", so they can not longer be marked as actual answers, even if they are the correct answer.

3) Answers consisting mostly of links to other SO questions with a brief explanation get deleted because "they don't add value". I can't show you them because they got deleted (duh!). That other question/answer might have provided useful context and discussion, or it could actually be the good answer; besides the link itself is valuable as it links two related questions together... In their zeal moderatos actually delete useful information from the system.

4) Security questions overwhelmingly have dangerously wrong accepted answers. SO should admit the shortcoming and ban all security questions (other than pen-testing).

Basically, SO risks becoming another wikipedia with deletionists ruling the roost.

The following website doesn't inspire confidence, but without doubt, FileHelpers 2.0 is the most awesome .NET library out there for processing and creating CSV files: http://filehelpers.sourceforge.net/

Oh, and its free thanks to the hard work of Marcos Meli.

"SO should admit the shortcoming and ban all security questions (other than pen-testing)."

You can comment on answers and even suggest your own.

I wonder if the "popularity" of a language is related to the fact people are learning it.

More people could be developing in Java, but more people are learning JavaScript and therefore active on Stack Overflow.

I generally find I use Stack Overflow a lot lot more when I am picking something up for the first time.

Therefore I would assume JS is the language that is being most actively learnt.

Not necessarily IMO, considering it still does make sense to ask questions even "after" you've learned a language, as long as you are using a language/platform and continue to push yourself outside of your comfort zone, you are bound to have questions. Wouldn't it be boring otherwise?

Sure, but the number of questions you have in the first 3 months of the language are bound to be more than the number of questions in the second set of 3 months, and far more than after you've used the language for 3 years. Granted most of the questions you have when you're "younger" won't require you to actually ask as you can find answers, but I'd be surprised if the number of questions asked by a user is relatively constant over the first several years of having learned a language.

I tend to see SO questions as being rather simple, such as why doesn't this API work, or what's wrong with my Syntax or what does this stack trace mean.

When you are more advanced your questions become more architectural and harder to convey in text.

Perhaps that’s just me…

That also harmonizes with my experience learning a language.

I really don't think 2.2% increase in happiness during two studies is statistically significant. Especially given the user base increased by over 300% in that time so presumably the demographic will have changed. Still, clever use of information to promote their site. Its amazing how fast stack exchange has grown.

I think the lighthouse image for 'Good management', number 3 under 'What's important to programmers when evaluating a job opportunity', is spot on. Really conveys that good management should provide very clear direction.

On a side note, I'm not sure why jQuery is listed as a programming language.

Programming language...jquery...

Judging from those results I think that the original question was more "what if" rather than "what is" - particularly in the salary area.

In 10 years here in the Washington DC area I've never seen salaries that high, or employers that can actually deliver that.

While I appreciate the information, the ad made the whole thing cheesy.

Why only where US developers live? What happened to the rest of the world?

Yep it would be interesting to see also international data (someone also posted a question about European data in the blog comments).

There are interesting stats on Alexa regarding SO (don't know however how reliable they are): http://www.alexa.com/siteinfo/stackoverflow.com Click "Traffic stats", then "Stackoverflow.com’s Regional Traffic Ranks".

Interesting to see SO being 22nd most popular site in India while ranking 166th in the US, and only 363rd in China.

The graphic is more US Centric, but at the bottom they link to the raw data: https://www.surveymonkey.com/sr.aspx?sm=vU4rF_2bPVQaftSo1s69...

This breaks down some major countries/regions of the world nicely. (e.g. US 37.7%, 'Other Europe' 16.7%, UK 9.6%, India 7.6% on down)

Considering the answer is mostly a list of the most populous states, it's particularly not useful.

http://xkcd.com/1138/

why is Stack Overflow still IPv4 only? Are they hosting on Amazon?

Reply on news.ycombinator.com