Data Shows Women, Men Equal In Coding Challenges

Screen Shot 2016-10-22 at 3.37.56 PM
Coding IDE

CodeFights, a coding challenge and recruitment website for developers, has some interesting insights on how developers like to work. They also have some compelling data on which programmers are faster, which languages are most popular, and which gender is actually better at coding battles.

In algorithmic challenges, C++ devs are faster and more accurate than anyone else by a healthy margin (a lot of algorithm material is written with the language in mind, though). Python developers come in second, with JavaScript devs rounding out the top three.

Sublime Text is also a favorite IDE of CodeFights users, in use by three-fourths of developers on the platform. Second place goes to vim (with 17.6 percent), while emacs comes in third with 7.6 percent of users. Roughly two-thirds of us like autocomplete ‘on’ as a default (who doesn’t want help?), and a whopping 91.4 percent of developers prefer a dark IDE to a light one.

Worldwide, C++ is the most popular language on CodeFights, with 24 percent of us utilizing it. It has some competition, though: Python has a healthy 22 percent of users, and both Java and JavaScript are two percent off the mark (20 percent and 18 percent, respectively).

Stateside, those numbers flip entirely. JavaScript is most popular with 24 percent of people using it, with Java a close second at 23 percent. Python is behind both at 18 perent, while C++ has just under 15 percent of users interested.

Considering gender lines, CodeFight finds we’re just about equal. When going head-to-head in coding challenges, Men only win 54 percent of the time.


9 Responses to “Data Shows Women, Men Equal In Coding Challenges”

  1. Byzantine_General


    if (0.54 == 0.46) {advance_narrative();}

    50 years ago I had to give up on explaining to a programmerette that she was never, ever, going to get 10 decimal digits of precision out of a 32-bit float. “Not that many 32-bit values” fell on deaf ears.

    Naturally, she was promoted into management.

  2. Programmerette

    This just gave me a bit of a laugh, seeing as I explained the problem with floating points to a male colleague of mine just today. Fortunately, I was articulate enough to explain it without issues, despite my weak, womanly brain! Turns out explaining exactly what the mantissa and exponent do is more helpful than repeating that it won’t work.

  3. Byzantine_General

    Hell, my Mom was exceptionally good at math. She was the first woman in her country to study college level math.

    Using Bayes’ theorem, she convinced me that she was luckier to be so gifted than I was, simply because of our respective sexes and the observed distribution of mathematical talent.

    And she also convinced me that the a posteriori likelihood of that which is the case is 1.0, whatever might contribute to the a priori estimate.

    The poor oppressed thing suffered from false consciousness. Nobody had ever told her that if you assign individuals to classes and then analyze the classes under the axiom that any difference between classes is due to “oppression” then the classifier owns the argument, because any desired conclusion can be reached by adroit choice of classification. Of course she was more interested in actual proof than social bullying, so there’s that.

    The nonexistence proof* of a bijection between {0,1}^32 and {0..9}^10 is independent of any encoding tricks, mantissas, exponents, whatevs.

    *There are more members in one set than the other. No bijection is possible.

    And /that/ was the point I was unable to convey yea these many years gone. Robin was convinced that there was a clever enough way to do it. Nothing I could say would budge her.

    Some people get rigor. Some don’t. The M/F distribution of that attribute is profoundly uninteresting.

  4. Gargarito

    I think the data sample in this article is flawed. Can you please provide a reference to the source of the data?

    For one, per other sites, PHP may be the most widely used language as it accounts for around 80% of the websites in existence. See:

    Another site, module counts ( which lists the “packaged” libraries for a language, lists number one among the languages npm, which is javascript based and number 2 java. Per this perspective, PHP, which includes packages via drupal, packagist and pear repositories still accounts for a lot.

    C++ does not even exist in the modulecounts site. C++ is very relevant, but is not very used in the web. It was used in drivers and apps a decade ago. Yes, games and other endeavors that possibly started a decade back are still in C++, but there is no shared repository that I know of, other than the traditional C libraries distribution method, which is hardly as scalable as the repositories method. Possibly nuget, which is .NET is the only repository which may have some C++ .NET packages.

  5. The core of most applications is written by a very small group of people The vast majority of “programmers” are only involved in maintaining and modifying existing software. That is why these articles about which is the “best” language are rather silly.

  6. Jack Jackson

    All these flash-in-the-pan technologies will come and go. However a hundred years from now we’ll still be writing and supporting Cobol applications.

    I can see it now some kid in the future going for an interview with Cobol 74 on his resume and being asked if that was the 1974 version or the 2074 version.