Follow Slashdot stories on Twitter


Forgot your password?
Stats Programming

Are There More Developers Than We Think? ( 191

JavaScript's npm package manager reports 4 million users, doubling every year, leading to an interesting question from tech industry analyst James Governor: Just how many developers are there out there? GitHub is very well placed to know, given it's where (so much) of that development happens today. It has telemetry-based numbers, with their own skew of course, but based on usage rather than surveys or estimates. According to GitHub CEO Chris Wanstrath, "We see 20 million professional devs in the world as an estimate, from research companies. Well we have 21 million [active] users -- we can't have more users than the entire industry"...

If Github has 21 million active users, Wanstrath is right that current estimates of the size of the developer population must be far too low... Are we under-counting China, for example, given its firewalls? India continues to crank out developers at an astonishing rate. Meanwhile Africa is set for crazy growth too... You certainly can't just count computer science graduates or software industry employees anymore. These days you can't even be an astronomer without learning code, and that's going to be true of all scientific disciplines.

The analyst attributes the increasing number of developers to "the availability, accessibility and affordability of tools and learning," adding "It's pretty amazing to think that GitHub hit 5 million users in 2012, and is now at 20 million." As for the total number of all developers, he offers his own estimate at the end of the essay. "My wild assed guess would be more like 35 million."
This discussion has been archived. No new comments can be posted.

Are There More Developers Than We Think?

Comments Filter:
  • no (Score:2, Insightful)

    by Anonymous Coward

    only if you call all js-monkeys as developers

    • Re: (Score:3, Interesting)

      This is actually an insightful comment. The industry started going downhill when we stopped having Software Engineers and Programmers and started calling everyone a "developer." Now any idiot can create a github account and start programming, with no idea how to design systems, call themselves a "developer", and get hired by a clueless "hiring manager" who granted the interview because his resume had the right buzz words on it.

      The vast majority of people writing code these days don't have a clue about met
      • Re:no (Score:4, Interesting)

        by Anonymous Brave Guy ( 457657 ) on Monday May 29, 2017 @08:42AM (#54504863)

        Not that I think you're wrong, but is this really so different to the Visual Basic or PHP "developers" of yesteryear? Maybe JS is just the latest language that everyone automatically has on their computers and where it's easy to find beginner tutorials (good, bad or otherwise) online.

        • The industry started going downhill around then ... I was not implying the industry started going downhill recently.
        • Same old same old. There were a bunch of kewl video games written in the 1980s Busing basic. And there were a shitload of piss-poor games written in Basic in the 1980s but no one remembers those.

          I've heard folks complain about the current writing skills of Americans saying Thomas Jefferson was a good writer but folks today aren't. Well TJ was one of the best writers of his generation. Most any luser here on slashdot can write better than the average human of 1760.

          90% of everything is crap.
        • I think the difference is in degree. We are in the middle of a groundswell in the number of coders, both in the US and worldwide. The lure of being the next Gates or Jobs (and they probably think he was a developer) has been sinking into society for a couple decades now. Add to that all the pushes to put disinterested kids into coding courses, or even to make it a mandatory course. All this is going to decrease the skill of the average coder, as people start getting into it for the wrong reasons.
          More low-s
          • While I agree with most of your premise...the one crack in its foundation is that coding isn't required for non IT jobs. Many many many engineering and scientific jobs are going to need passable coding skills to be competent.

            Even within the IT world, the 'tester' is fast becoming their own flavor of a programmer. It's no longer viable to manually plod through an application as a human. You need repeatable testing 'scripts' that exercise that application. That's programming as well.
        • by murdocj ( 543661 )

          Sadly, JS is far, far worse than Visual Basic... at least VB 6, which I programmed in many years ago. That language actually let you build reasonably structured programs, despite having the "VB" moniker. JS is just an invitation to disaster.

      • by mark-t ( 151149 )

        Now any idiot can create a github account and start programming, with no idea how to design systems, call themselves a "developer", and get hired by a clueless "hiring manager" who granted the interview because his resume had the right buzz words on it.

        ... And then get themselves fired within only a few weeks of starting as soon as the CTO and programmer lead on that person's team quickly realize that the guy doesn't know jack shit. I've seen it happen. Obviously there's some cost to the company for doi

      • by mccalli ( 323026 )
        This is rose-tinted thinking. We didn't have a clue about methodologies and processes in the 80s either, when I taught myself to code in BASIC. Version control? I'd tell you I'd heard of it but I hadn't. Early 90s I did - I copied directories to different names. rcs etc.

        There's irony in putting that in an article praising GitHub, i.e. git, written by a guy who resisted source control for so long and then claimed it was because every system was rubbish except his. He knew nothing about process either.
        • You mean you didn't have a clue about methodologies and processes in the 80s. We certainly did [].
        • Directories?!!?!? My version control in the early 1980s was a set of numbered cassette tapes. Directories were for the rich people who could afford floppy drives.

        • git, written by a guy who resisted source control for so long and then claimed it was because every system was rubbish except his

          You mean, Linus Torvalds wrote it after bitkeeper stopped providing free hosting for the linux kernel.

      • It's always been that way. 45 years ago it was a bunch of COBOL programmers who couldn't write a linked-list to save their life. There have always been a core of a few programmers who are really good, with a huge mass who spend their lives implementing business logic, and never learn anything else. Even in the 50s when you had to stick wires into plugboards, there were a lot of people who didn't know what they were doing.

        (Implementing business logic in itself isn't necessarily bad, it pays the bills).
        • COBOL programmers actually couldn't write linked lists 45 years ago, because COBOL didn't have pointers at the time. It certainly has not always been like this by any stretch of the imagination. Back then simply being able to write software meant you were much smarter than average. Today the boy and girl scouts literally have teenagers thinking anyone can do it (for example.)
          • simply being able to write software meant you were much smarter than average.

            That's definitely not true. IBM tried at one time to hire programmers based on chess ability, but that was definitely not widespread. When Knuth wrote AoCP some of his readers were shocked at the idea that they could write their own data structures and algorithms for common things, instead of using the vendor supplied packages.

            • Chess ability is no indication of ability to program, and I'm sorry to break it to you but there were no people of average intelligence allowed anywhere near a computer originally.
              • there were no people of average intelligence allowed anywhere near a computer originally.

                Why do you even think that? Do you have some kind of research to demonstrate it? We're talking about the late 50s and 60s btw, not when the only computers in existence were at the Institute for Advanced Study or something. Despite what you may believe, not everyone in those days was like Mel [].

                Seriously, look at this code and tell me that it was good []. I've seen ugly Javascript but never that bad.

            • Chess is far too boring for me. Reversi: Now there's a game!
          • I've implemented linked lists in DOS BASIC. If you think it requires are what this article is about.
            • There is no such thing as DOS BASIC idiot. Now once again, off you go little turd ...
              • you're still the one who thinks you need pointers to implement a linked list.
                • You have to be the dumbest motherfucker on Slashdot today. Linked Lists [] involve pointers. If the language doesn't have a built in facility to accommodate pointers then you have to use built in facilities to implement them, which is slow and phenomenally stupid. In short, you can do almost anything with a Turning complete language, except implement a Linked List without pointers. Now just shut the fuck up. Every time you post you make yourself look more and more stupid.
                  • A 'pointer' is whatever construct you use to point to the next record in the list.

                    From your link:

                    It is a data structure consisting of a group of nodes which together represent a sequence. Under the simplest form, each node is composed of data and a reference (in other words, a link) to the next node in the sequence.

                    Data and a reference. If your data record contains a record number of the NEXT record in the sequence. That's a linked list.

                    [[ a ][ 1 ]] [[ b ][ 4 ]] [[ c ][ ]] [[ d ][ 3 ]] [[ e ][ 5 ]]

                    Gives you A -> B -> D -> E -> C

          • You can implement linked lists in arrays. You don't NEED pointers.

            • You used pointers idiot. You didn't use built in language facilities that support them, so you used the built in facilities to create them. ALL linked lists use pointers. It is LITERALLY the definition of a linked list.
              • And thus, you can implement a linked list in COBOL, can't you?

                Reminds me of the Advanced C class I took with the following problem:

                Const char[] strIn="Find the comma, and change it to a period please";

                My methodology- grab the pointer to strIn, increment it until I found the comma, and change it to a period (I leave it as an exercise for the reader to implement).

                Everybody else used a strcpy first, because after all, it was a constant.

                My score? -5, scribbled out, and a note "I tried it, it worked".

                What I kne

      • by kriston ( 7886 )

        Sad but true.

        There aren't 35 million npmjs developers. There are 35 million npmjs uploaders. That's all. This is a non-story.

    • Re:no (Score:5, Funny)

      by Anonymous Brave Guy ( 457657 ) on Monday May 29, 2017 @08:32AM (#54504837)

      If there were 5 million in 2012 and 20 million in 2017, that's probably only a million JS programmers, they just created new names for everything every few months...

      • Re:no (Score:4, Insightful)

        by TheRaven64 ( 641858 ) on Monday May 29, 2017 @08:40AM (#54504857) Journal
        More seriously, there are lots of reasons to have multiple GitHub accounts. There are quite a few people at Microsoft, for example, who have an account with MSFT in their name for open source work-related stuff and another account for private stuff. Additionally, GitHub uses a horrible ACL mechanism, rather than a simple capability mechanism, for authorising applications, which means that the easiest way of granting rights to monitor a repository to something like Travis-CI or Coverity or similar is to create a new account that has access only to that repository and create an OAuth token for that user. It's not surprising that the number would quadruple when there are lots of good reasons for people to have 2-3 accounts...
        • I always wondered about this. Unless something changed, I thought that GitHub's official policy was that one person is only allowed one (non-paid) account, possibly with some exceptions for things like well-known projects. But IME there are mostly only two kinds of developer: those who don't really use GitHub much other than anonymously pulling from the repos, and those who are really into it and often do seem to have several different accounts they use.

        • I admit I personally have two accounts.
        • It reminds me of when Windows Phone 7 came out. Instead of releasing sales numbers, Microsoft released the number of downloads of its SDK, which was around 10million IIRC. At that time, I had already personally downloaded the SDK two or three times for work.

          In the end, I think there were more SDK downloads than phones sold. Again, that is certainly true in my own case, I didn't buy a single one.
        • by Xest ( 935314 )

          Not to mention the IT ops side of devops usually have accounts nowadays because they often get involved in building and deploying such software even if they never write a single line of code, and with most large companies now going the devops route, you're probably looking at anything from hundreds of thousands to a few million users there alone.

          I actually typically have 3 accounts for this sort of thing - day job account, personal account, and an account for the startup I'm helping in my spare time, so not

    • You're right to say "no", but for other reasons. For instance, many apps require that their users create Github accounts in order to submit bug reports and feature requests. If you look through issues on user-facing apps, particularly smaller ones with no other channels for submitting feedback, it's usually pretty evident that the majority of people have no background in programming.

      Then there are the students and hobbyists who create an account to work on things on a rare basis. They may be in the pool of

    • only if you call all js-monkeys as developers

      I mean, I'd never have typed that comment, but it's what I was thinking.

      I know way too many js people.

  • Humanities too (Score:2, Interesting)

    by Anonymous Coward

    These days you can't even be an astronomer without learning code, and that's going to be true of all scientific disciplines.

    My cousin has to develop a lot of customized software as an economist.

    Here's something else - historians.

    Historians have been using modern imaging and are digitizing old documents. For example English church records. And using "big data" techniques they have been discovering new things about history.

    Although, people like my cousin find the coding a tedious chore that they have to do to solve their problems and dreams of the day when we have computers like on Star Trek: "Computer, what is the relationship be

    • I think we should be more critical about what we call developers. This is not to belittle the many, very competent, non-professional coders, but saying that everybody who downloads a toolset and writes code is a developer is a bit like calling everybody who has bought an electric powersaw a carpenter; there is just more to being a professional developer than writing code, and there is a big difference between writing short pieces of ad-hoc code from time to time and writing large, maintainable applications

  • by Anonymous Coward

    Well we have 21 million [active] users -- we can't have more users than the entire industry

    It's impossible for a single user to register multiple accounts. Impossible, I tell you!!!

    • by tommeke100 ( 755660 ) on Monday May 29, 2017 @08:40AM (#54504855)
      How do they define "active" users? I have a GitHub account but rarely use it, mostly for some open coursework machine learning examples / tests and Kaggle now and then. I am a professional developer though. I wonder how many GitHub account are just plain students setting one up because they need to for some course, but aren't developers.
    • by Tukz ( 664339 )

      And not all those users are developers.

      So yes, they CAN have more users than there is developers.

    • Also, he doesn't understand why people besides "professional developers" might have GitHub accounts.

      • I'm a professional developer and have no git hub account.
        Why would I?
        I don't host my professional code on a server out in the internet.
        I'm not even legally allowed to do so. All companies I work for have their own git or svn servers.
        The few hobby projects I work on I host on my git and on my back ups ... there is no point in having a git hub account.

        • by green1 ( 322787 )

          Meanwhile I would never describe myself as a developer, but I do have a github account.
          I tinker with the occasional thing at home, an rpi for this or that, or my own server that I "play" with. Having a few web pages, and a bit of home automation does not make me a "developer" but I might be counted in this.

          I think they need to realize what while there is probably some overlap between developers and account holders, neither one is a good indication of the actual quantity of the other.

        • Shorter angel'o'sphere: "Stop liking what I don't like!"

          Just because you don't see a point to having a GitHub account doesn't mean there is no point for anyone.

          • That was not the point :D

            The point was: you can not calculate the amount of developers world wide by looking at git hub accounts, or bit bucket (there I have an account, and the username is not angelosphere)

        • by pjt33 ( 739471 )

          There are reasons to have a GitHub account even if you don't host any code there. The obvious one is to be able to post issues on projects which are dependencies for your projects; slightly less obviously, is in some senses the new Geocities.

  • In case of the article a developer is anybody who writes code. In that context, 20 mio is far too low. However, in case you want to count only people who actually write code for a living then these number would go down. Still I think 20 mio is a little low on that, as we have 7000 mio people on the planet.

    • by green1 ( 322787 )

      "Anybody who writes code"? is that really a sensible metric though?
      Does the purpose or quantity of that code count?

      Compare to other industries. If I make myself a sandwich at home and I a chef? If I replace the faucet in my bathroom am I a plumber?

      If I have my own tiny webpage and an rpi that I tinker with in my spare time, am I a developer?

      I would never call myself a developer, any more than I would call myself a chef or a plumber, but I have done all of the above. I have a github account, so I guess I'm s

  • Ok so... (Score:5, Insightful)

    by CODiNE ( 27417 ) on Monday May 29, 2017 @08:05AM (#54504789) Homepage

    If I doodle am I an artist?
    If I put air in my tires am I a mechanic?
    If I floss an I a dentist?
    If I buy plants am I a horticulturist?

    It seems if merely downloading some code makes one a developer... we have a serious respectability problem as a profession.

    • by green1 ( 322787 )


      While there's probably some correlation between people who are developers, and people who have github accounts, I'm quite certain that it's not anywhere near accurate enough to get any sense for the quantity of developers.

      They've done the equivalent of counting the number of people who buy groceries at one of the major food chains and assuming that they're all chefs.

    • It seems if merely downloading some code makes one a developer...

      Nope. But modifying it does. It doesn't mean anyone should hire you, mind you. Getting paid makes you a professional, it doesn't make you a whatever. You can be a hobbyist whatever just by doing whatever. That's all it takes.

  • by Anonymous Coward on Monday May 29, 2017 @08:06AM (#54504791)

    There is a typical logical disconnect in this summary. A lot of GitHub accounts are for students and people who are just hobbyists. Therefore a GitHub account doesn't equate to a professional developer. Also, multiple GitHub accounts per person is not abnormal.

    The researchers indicate that there are 20 million professional developers. In other words, not including students and hobbyists.

    Therefore the GitHub CEO is a moron, or his statement is out of context or is mis-quoted. Or the submitter/story writer is just making stuff up.

    20 million professional developers globally seems a reasonable estimate. Somewhere above 40 million people globally who code regularly is not unrealistic.

    • I think I'm the only person in my group with a GitHub account. Nothing we develop is opensource. None of our compilers are opensource. (NXP just released a VLE enabled GCC).

      At my company, competitors, and others in industry there are easily millions of embedded developers that aren't being counted.

    • 20 million professional developers globally[??] seems a reasonable estimate.
      Reasonable estimate? As git hub clients/customers. Perhaps. As anything else, no.
      According to Stackoverflow Germany has 700,000 software developers from about 80,000,000 inhabitants and 40,000,000 work force.
      I would imagine that scales to far more than 20M developers world wide.

  • by Anonymous Coward on Monday May 29, 2017 @08:28AM (#54504823)

    > we can't have more users than the entire industry

    Sure you can. I personally have 3 different github accounts created with different email addresses. You?

    • by asylumx ( 881307 )
      Ya, this is the most obvious counter-example. Seems pretty elementary to assume that many users have multiple accounts, perhaps at least one for work and one for personal use. Also really obvious that people who aren't developers can & do have accounts (whether you count hobbyists or not) and that tons of professional developers don't use GitHub.
  • by grahamtriggs ( 572707 ) on Monday May 29, 2017 @08:36AM (#54504843)

    And certainly not all users are coders (depending on quite how you term "developer")

    Yes, I am a developer, but amongst the activities I use GitHub for, I host a website. Not all of the contributors (who have user accounts and have submitted pull requests), are developers.

    What about people just uploading data sets to GitHub for sharing?

    What about people that contribute just to the artwork or documentation of a project (where those files are in GitHub)?

    What about users who have an account solely to open issues in the issue tracker? Or contribute to wikis?

    And then there are students, or even just hobbyist coders, never in the industry but just doing it for fun?

    Saying that you can't have more users than the industry is pretty dumb, and suggests that the CEO doesn't understand his own product.

  • I was talking with my cousin over the weekend and he thought programming was just sitting at a computer inputing data from spreadsheets.
    The hard truth is, developers are sitting in front of a computer making things happen. It can't be that hard cause they're just sitting in front of a computer, right?

  • Are the purposefully writing to enforce Betteridge's law?
  • In addition to the true observation that not everyone with an active GitHub account is a qualified software developer, it ought to be pointed out that the promotion paths most companies have discourage developers. When people get good at it, they tend to have the option of having a stagnating salary or moving into management. Other developers move sideways to get more lucrative positions in sales-related jobs. Both of these work to pull qualified developers out of the pool. Many of these former developers s
  • This depends on what we define as "developers".

    Are we talking about people who have had formal comp sci education? Or people who read an html tutorial once and thought, "Oh yeah this is easy! I am a developer now!"

    The number of people I've run into who think they are gods gift to software development, but don't actually know WTF they're doing, is staggering. Worse is when these guys have just enough charisma and knowledge to bullshit their way through interviews with people who don't actually know better

  • between those who willingly look at any code at all, and those who won't.

    Yes there are a lot of bad or unprofessional programmers, but in terms of general thinking ability the bar goes so much lower.

    Go ahead and count them.
  • How can any stats at GitHub give you a clue about the total number of developers, unless you just happen to be the one person in the entire world who, magically, has a vague idea of what fraction of developers use GitHub?

    What fraction of developers use GitHub? 0.1%? 1%? 10%? 31.6%? If you pull 20% out of your ass (and that's exactly where you would be finding this number) and just to play contrarian, I say "No, you're wrong, it's only 10%" or "you idiot, it's more like 40%" can you explain how closer to ri

I don't want to achieve immortality through my work. I want to achieve immortality through not dying. -- Woody Allen