The Stack Overflow Developer Survey 2018

Every year our portfolio company Stack Overflow surveys its developer community and publishes the results.

This year Stack had over 100,000 respondents to its survey from all over the world, making this survey possibly the most comprehensive view of the global software developer community.

There is a ton of data here. It’s a 30 minute survey. You can see the results here.

But since many of you won’t click that link, here are some highlights from it:

First, we know that software engineering is a largely white male profession. The data shows that:

If we look at the gender and racial/ethnic mix of the students who answered the question, there is some promising data on racial/ethnic diversity, but less promising data on gender diversity. Efforts like I blogged about yesterday are badly needed to change these numbers.

I found the technology questions interesting.

Javascript is by far the most common programming language.

But Python is the most “wanted” programming language. And Go and Kotlin are rising fast.

Some great news for our portfolio company MongoDB in this survey. Mongo was the most popular non-SQL database and was the most wanted database of them all.

Finally, some data on how important Stack Overflow is in these developers’ work:

Any service where 2/3 of its users visit daily is a big deal. And for developers, Stack is very much that.


Comments (Archived):

  1. PhilipSugar

    I think the key is starting early. I believe that kids look up to their parent of the same gender. I.e. girls/moms boys/dads.Now what this means is if we have a gap now that gap will continue unless you break it.The only time to really break it is early. Because by the time you finish college, it’s pretty much (not totally) over. And that everybody can work on. I think the work that Fred is doing in schools is great.But you as a company can do things like sponsor and mentor a lego league programming team (we won best programming in the state)You can go to the school and ask the TAG (Talented and Gifted) or whatever they call it by you teacher to see if they can get some good kids including girls to join (that is what we did)But you do see the bias, and that is where you can intervene. We would see the girls say: “I’ll do the poster and skit”. No, you are really good at programming everybody does everything evenly. And you know what? They were the best programmers.

    1. LE

      Exposure and influence are for sure important. But as I have found out with my stepkids some people just aren’t cut out for what you would like them to do. My stepson math wiz (current addiction Fortnight) zero interest in programming. My god he could be good at it but not even close to considering doing so. His father (my wife’s ex) buys him a heped up gaming computer. To lazy to even take it out of the box. Has been in his room for perhaps a year at least. That’s lazy. He gets enough of a game fix from a small laptop and doesn’t need anything ‘better’. Maybe @falicon will buy it or I will just take it one day and give it to someone who can.Otoh my stepdaughter type of person that will go downstairs with me and help change the HVAC filter. And comes to my office to do work without pay. Just to learn what goes on. She did lego league for several years as well. Other kids on the team were primarily children of Indian immigrants. Those kids are kicking ass and comprise most of her friends.My two daughters (older with jobs in NYC now) were traditional girls. Even though they were raised by my ex wife no way would they have any interest in programming and it has nothing to do with lack of exposure either. (My ex wife is married to someone who is world known as head of this or that at Adobe..)Edit: Lest I appear to be slamming my stepson my frustration comes from the fact that he is so highly talented and ranked top at his school but doesn’t want to put in any effort given what he could be doing with his time. I think this is like a problem for ages with parents actually as in ‘wow knowing what I know this is what I would have done if I was you and had your ability’.

      1. PhilipSugar

        Yes, you cannot force people into things. Including making somebody work.It is interesting that yes, the Asian and Indian Girls are very good and again, that confirms my point because both of their Mom’s are technical. That is the only way you know you are going to get to come to the U.S. (don’t get me started) We eventually had to ban Mandarin getting spoken in the office, except to parents as two of our kids would get into bitter arguments but we couldn’t understand.

        1. LE

          It’s kind of funny how you end up with this positive view of a particular culture (Indian in this case) as a result of being exposed to the type of people (out of a country of 1 billion) that are special enough to make it here. Or to want to come to the US. Some of these kids are talking about going to Oxford and they are only in middle school. Growing up in an immigrant family myself where all that mattered was work and education (no play no joking) I can definitely see the advantage (in brainwashing) that they have over today’s pampered American families. [1] No wonder we are getting our asses kicked and need to admit bell curve immigrants.[1] Of course I can counter my point easily by how my sisters think vs. how I think even though we were raised in the same family.

          1. sigmaalgebra

            See my post to JLM where I explain US women based on a 1951 movie. Then, to explain your point, it’s not just the upper tail of a bell curve or “pampered” US women but the desire of US women to fit in with their gossip group.A related point is, 100+ years ago, wives were at HOME. They were nearly all farm wives, and there was plenty to do. Now for nearly all that home labor, there are appliances, products, etc. Well, these need to be paid for, but not a lot of young women had mothers and grandmothers who did well in the world of work. Besides, the norms have not yet fully changed.

      2. sigmaalgebra

        Let him take a look at… Neil Turok Public Lecture: The Astonishing Simplicity of Everything October 7, 2015Uh, the stuff on the blackboard is not like what is in the lecture.

      3. JamesHRH

        Having that ability is, in reality, an impediment to working hard.When you can do things easily, the hard way seems dumb.

    2. fredwilson

      That’s why I have made it my primary effort outside of USV to get CS taught in all of the NYC public school buildings, especially the elementary schools

      1. PhilipSugar

        We agree 100%. Your effort is much bigger than mine, but I believe every bit counts, and if we get 10,000 companies putting in an effort, that will add up.

  2. JLM

    .I am amazed at the gender disparity in ALL and STUDENTS. In a million years, I would never have considered those percentages possible. They are overwhelming and staggering.It is clear the disparity will continue when one compares the present (ALL) and the future (STUDENTS).Wow!This is a very useful document and one which conveys real information.Well played.JLMwww.themusingsofthebigredca…

    1. LE

      It is a stack survey of people who use stack. While statistically it is probably correct it would be interesting to compare the results to a control group. For one not everyone who even uses stack answers survey. Just like not everyone who reads AVC makes comments. And not everyone even will take a survey. I never do surveys. So you have a group of people answering and perhaps there is a difference between men and women and whether they take surveys.Now that said do I think anything that I said will change the results a great deal? No. I think it’s probably very accurate. But just the same survey design and the group of people being surveyed can greatly change the results (as you know).

      1. JLM

        .Agreeing with you more than you do with yourself, I am still stunned at the gender breakdown.JLMwww.themusingsofthebigredca…

      2. PhilipSugar

        I was going to bring that up. If you work for one of the big chemical or banks near me, you probably aren’t even allowed to use it, That also will cause some bias on Microsoft and Oracle.

        1. LE

          Yep. Also the idea that you have people working for you and you allow them to share the secret sauce of how they do what they do is honestly quite ridiculous. On your time! [1] It is something born of the internet age. Can you imagine someone at Walmart back in the 70’s giving away the finer points of how they operate? “Hi today I will do a blog article explaining exactly how we pick our site locations. Now I know my larger competitors already know this. But I figured I would tell anyway so some new company can use the information and perhaps grow into a competitor!” Or in sports “here is how we get the edge good luck and write me if anything isn’t clear or what we do!”You know why they went down this path? Because there was so much business it didn’t matter. Try being in a traditional business where you are locked in a death match for sales with competitors and see how generous you are. You just stuck out your hand and business came. No beating your head against the wall.[1] Or as we used to be able to say ‘retarded’ (used for editorial impact).

          1. JLM

            .Actually, Walmart – which looks at a million deals – lays out its selection criteria as to the demographics of desirable sites as a screen to keep people from submitting bad sites.They are very open about this.Any site has to be run through an elaborate demographic model – all available on the Web now – to determine population density, income, buying power and competition.This was novel 30 years ago, but not now.You can lookup the average sales per $$$ of income given the actual demographics. Higher income, obviously, indicates higher sales. They have to be “Walmart” customers which you can determine from the same demographic and income info.Then, you get the sales tax info on all Walmart competitors within 10 miles and see how much of the perceived demand is already being fulfilled by current offerings.When/if there is a void — more income and buying power — you have a shot at a deal.The last step is to take a look at traffic at the site and all competitors.I did one deal with Walmart and found them to be very professional and forthcoming. They are also very cost conscious and frugal, willing to pre-invest in growing traffic over a 20 year time period.Walmart can, effectively, access capital at lower than the US gov’t. When they really like a location, they will build their own Supercenter across the street even when they have to pay out a long term lease.There is a whole world of real estate trying to figure out what to do with former Walmart buildings.JLMwww.themusingsofthebigredca…

          2. LE

            Crap for sure that was not the best example on my part. (Using real estate selection). I should have picked something more abstract like arranging things on shelves or end caps (that of course you can figure out in many ways (including reverse engineering)). To my point you can go into a walmart and take some pictures but you would be flagged if to obvious and spent days there filming everything. Point is no positive benefit by educated future competitors even though in theory someone can work for you and leave with the same knowledge.My cousin (mr. supermarket) used to tell me he could tell how much business a competitor was doing by simply getting the milk delivery guy to tell him the amount of milk they were selling.That said one thing I will add that there is this common misconception (and it doesn’t detract from what you are saying) that ‘that info is already out there’ like in ‘if a thief wants to break into your house then locks won’t stop them’. There are many people that get spurred on by reading and the more info and help out there you never know when someone will come upon it and act upon it even if it is available if you look for it or know it’s out there. So the easier it is or more widely known that doesn’t work in your favor.I

          3. PhilipSugar

            I would make two points. Walmart is willing to give info on how they are going to pick a store. Benefits them.If you think you are getting their logistics algorithms or sales product placement (how they put stuff on shelves) you are crazy. I know this for a fact. If you as a software vendor go on a pitch the room will be full of other developers. That is how you know you are screwed. Fedex, UPS, HomeDepot, Lowes, and others round out that group.

          4. Salt Shaker

            Are retail analytics and its inherent skills transferable to the digital world? That’s WMT’s challenge. Stock tanked on modest digital growth last qtr., but I’m long on WMT. The notion the market isn’t big enough for 1-2 players beyond AMZN defies logic. Plus WMT and WF’s footprint is dramatically different.

          5. JLM

            .I was in the Whole Foods Mothership in ATX last night and for the first time saw a display of all Amazon digital products – Echo, etc.You can begin to see the Amazon fingerprints on the tattoos, piercings, dyed hair of WF.I never saw the joint so clean and well stocked. There is a new step in their walk.JLMwww.themusingsofthebigredca…

          6. Salt Shaker

            It’s a great store, and always has been. Pricing was the killer, plus I never found it a particularly relaxing place to shop. Always felt rushed to get in and get out. Great quality products, beautifully merchandised. I’ve noticed a lot more emphasis on their P/L brands, with competitive pricing on “specials.” Let’s face it, AMZN provides WF w/ ground cover, margins be damned. Will be interesting to see how/if they breakout WF’s financials.

          7. JLM

            .Under GAAP, as a “unit” for the evaluation of goodwill, they will have to break out WF financials though not in much detail.I assume, based on the acquisition numbers, their is goodwill to evaluate — goodwill being the arithmetic difference between acquisition cost and depreciated book value of hard assets.If they buy something else which is considered “like kind” then they can elect to consolidate them into a single unit for goodwill evaluation purposes.JLM

          8. PhilipSugar

            Agreed on all fronts. I’d add the customer base is too. I know this sounds strange, but many Walmart customers do not have credit cards and go there on payday (not necessarily to cash checks because they have money)

      3. Salt Shaker

        Stack Overflow’s survey isn’t representative of pop at large or all those who code, and I’m willing to bet even the size/gender of the Student cell isn’t accurate. Usage/access to SO likely varies by level of sophistication and where one is in the CS learning curve. There likely is a standard bell curve w/ respect to Student coding (light to heavy experience), with Stack’s value prop commensurate w/ one’s knowledge and experience. Additional segmentation of the data would be helpful.

    2. PhilipSugar

      That is what they are, in my 25 years of experience. We used to be a bit higher at 20% but we are back down near 10%

    3. Salt Shaker

      Is this a marketing and/or product problem? The gender data between ALL and STUDENTS indeed is startling, but it suggests lack of demand, interest and/or awareness. If there’s a conscious goal to change perception among students, a worthwhile cause, the data suggests there’s quite an uphill climb. Given the CS/STEM initiatives in cities like NYC, it would be interesting to see if there are major market vs. rural skews to the gender data, assuming the sample size holds up as one drills down.

      1. sigmaalgebra

        Looks like the hacker bro culture took its toll.

    4. sigmaalgebra

      For the potential of females in high end science, as athttps://www.quantamagazine….look at the career ofYvonne Choquet-Bruhat.For an explanation for the US, look at Father’s Little Dividend (1951)Spencer Tracy, Elizabeth Taylorat…and notice the interests, proclivities, thinking, concerns, interactions, emotions of all three of the married women in that movie.Back in 1951, US movie audiences were fully prepared to accept those women as realistic.Congratulations both to the writers who saw such general patterns in women and understood the patterns and women well enough to get them both clear enough in a screenplay and to the actresses and director who got the same up on the screen so that audiences could see and understand.So, that movie is a good view of where, with some exaggeration, norms for US women commonly were back then in 1951. Then, no doubt a lot of that is still with us. What was it, “The apple doesn’t fall far from the tree”?It’s not that those women really have to be just that way. They can be different. But it is heavily that they want to “fit in”. Or, from the crib, human females are interested in people and, males, things. And, then, for life, the females are desperate to be surrounded by women and children and hate to be alone. So, when women are with other women, all the women want to fit in. So, quickly the norms of the group of women move to an average, and that has to do with people, not things. It appears, as in the movie that a lot of women are not even very comfortable being with only their sons and husband. To repeat, women want to be with women and children, and that’s really strong stuff.As in the movie, the women LOVE to gossip. Exchanging gossip is much of how they get position in the group. Well, they will want to gossip especially about women and children, marriages, pregnancies, babies, recipes, home decorating, fashion.About C++, C#, Python, JavaScript, …? NOPE!!! Eigenvalues and eigenvectors of a symmetric, positive definite matrix and their role in data compression, classification, and quantum mechanics? Not very likely!In the movie, Tracy’s wife is nice, BUT: Tracy gets motivated and spends an afternoon, his first in years, in a gym. Then the next day he is stiffer than raw spaghetti and really sore. So, in the morning, his sweet wife gets all concerned but makes four mistakes: (1) Doesn’t appreciate that all he has are some sore muscles. (2) Wants to call a doctor — absurd for just sore muscles. (3) Wants to call their new son in law — even more absurd. (4) Conjectures that the problem might be his appendix which she knows he doesn’t have and which would give very different symptoms anyway. The writers got a great list of four!!!! Sweet? Yes. Rational? No. Definitely ditsy.One of the iconic, telling scenes in the movie is the baby shower for Tracy’s daughter (Elizabeth Taylor, maybe 19 or so): So, the living room is packed with chattering, gossiping, posturing emotional, excited, really happy women, wrapping paper with pastel colors, tissue paper, baby clothes, etc., and Tracy observes that that room is no place for a man!!! He’s RIGHT.So far, in the US, women don’t want to be part of the hacker bro culture. For the world of work, women are happier doing what other women are already doing, K-12 teaching, nursing, HR, social work, clinical psychology, retail clerking, secretarial, law, medicine, reporters and interviewers on TV. E.g., on YouTube, look at the Right Side Broadcasting coverage of Trump’s PA speech and notice the two hosts interviewing members of the audience, a nice but socially awkward guy and a charming, animated, smiling woman! So, these careers all have to do with interacting with PEOPLE and often HELPING people. Did I mention that females are interested in people? Helping people? That helps them stay in good standing with the women in their gossip circle.Once my wife and I had a kitty cat, a really sweet cat. The three of us were walking someplace and at one point were walking across an open field, about the size of a football field. I was carrying our cat. Well, as we got to the center of the field, the cat became REALLY afraid, wanted to get down and run for the tall grass, trees, woods, etc. but get the hack OUT of an open field. The sweet thing even scratched me a little; I had to hold her more tightly.Well, there are a lot of situations in our society where men are comfortable but women are as afraid as our cat was in that open field.That’s just the way things are. Men need to know such things.Now, women can cook, sew, do bookkeeping and accounting. Well, then, they should also be able to do a lot of programming, especially for Web sites. Quickly they should be better at that than most of the men. Why better? Because the woman have better verbal talents, have better color sense, are better at clerical work, are much better at understanding the emotions of the users, and are better at being neat and organized. IMHO, the issue is not really fundamental ability but social norms. And the women want to fit in with the existing social norms.

  3. LE

    For anyone not aware our own @vasudevram… is a python expert…

    1. Vasudev Ram

      Thanks for the recommendation, @LE ! I appreciate it.

      1. sigmaalgebra

        Okay!!!! My main language for my startup is Microsoft’s Visual Basic .NET (don’t laugh; it’s essentially equivalent to C# but has a different flavor of syntactic sugar), but I do have some specialized, side issues for which I’d consider something better for those issues.Q 1. Why should I use Python instead of, say, Visual Basic .NET? For:(A) Screen scraping and/or HTML file parsing? For building and running a parser for a language defined by Bachus-Naur Form (BNF)?(B) As a scripting language for running other programs, say, in what used to be called batch?(C) For high level language, easy access to all of the old Windows System32 calls for writing old Windows graphical user interface applications?(D) For reading in data, say, comma separated values, name, type, value triples, XML, or something similar, drawing graphs and converting them to, say, GIF or PNG?Q 2. Since I’m programing on Windows, should I use Microsoft’s Iron Python which, IIRC, is compiled instead of interpretive, and has ready access to all of the .NET Framework software objects?Q 3. How would Python compare with Microsoft’s Power Shell in suitable uses, ease of use, features, capabilities (what it can do, not in the sense of security and access control lists).Q 4. How good is Python for writing programs to make use of Microsoft’s SQL Server?Q 5. Suppose want a much better program than the old Linux Grep or the Windows Findstr. l would Python be an easier way to write such? E.g., for some collection of text files, find all files that have in one or more paragraphs at least three of the following list of 15 words?Q 6. Can Python get to some lower level stuff, say, on what processes are running, virtual memory paging rate, I/Os per process, how real main memory has been allocated, what the TCP/IP stack has been doing, e.g., time outs, errors?

        1. falicon

          It’s almost all about syntax and structure…most all of the (popular) programming languages can do the things you list.Some make parts easier and some make parts harder, but all can probably do the things you are asking in one way or another.Ultimately it’s like saying “I can speak English, why should I learn Spanish or any other language?”…in the long run it’s not so much about what the language itself allows or empowers you to do, but rather who the language allows you to communicate and expand your knowledge with….so really it becomes a religious or a cultural argument more than anything.Personally, I think the answer is “whatever language you feel most comfortable and effective with, whatever one has a community you feel most akin to, is the one you should use”.

          1. sigmaalgebra

            Ah, but Python is supposed to have a lot of “packages” especially good for specialized purposes! I’ve heard that but know nothing about any of the packages!!!!!!!I will admit that Visual Basic .NET has “reflection”, permits looking at some aspects of the source code during execution!!!

          2. LE

            but rather who the language allows you to communicate and expand your knowledge with.Exactly. Same reason I went with Sun early in the decade and ditched SGI servers. The SGI’s outperformed the Suns. But the Suns were more widely in use.In many cases the brand halo and technical leader is inferior. But someone will offer up (some purist) some reason why it’s the best solution. But that ignores the stupidity of the market and how the masses actually buy and use.In consumer circles you don’t hear much from Leica (cameras) but Canon and Nikon are still ticking away (or even Sony). Ditto for Microsoft (god I still curse them) vs. Mac Osx (who I also now curse).

        2. Vasudev Ram

          You asked some interesting questions. I also see that you got a few replies. I am a bit busy, will answer in a day or so, but in brief, the answer to many of your questions is a yes, and in some cases a qualified yes. The answer is a no for very few questions, which is one of the beauties of Python – that is so widely applicable across domains, if not always being the best for all domains, at least it is often good or usable.

        3. Vasudev Ram

          Okay, here are some answers:>Q 1. Why should I use Python instead of, say, Visual Basic .NET? For:Not having worked on VB.NET much, I can’t say why you should use Python instead of VB.NET; I can only say whether Python is suitable (IMO) for the areas you asked about.>(A) Screen scraping and/or HTML file parsing? For building and running a parser for a language defined by Bachus-Naur Form (BNF)?See tools like requests, urllib3, Beautiful Soup, lxml, Scrapy.For writing parsers for languages (whether defined in Backus-Naur Form or not): see parsing libraries like PLY, PyParsing, etc. And Peter Norvig’s Lisp in Python:… (He is Director of Research at Google.)>(B) As a scripting language for running other programs, say, in what used to be called batch?Yes, Python can do that too, except that it is more oriented towards Unix, having been developed there initially. It still has somewhat good Windows support though. The work may not be as fluent as say writing your scripts directly in shell (bash) or in Perl but it can do it, either with custom code or with libraries. I had blogged some about this, but off the top of my head, plumbum is one, sh (a Python library, not the Unix shell) is another, osh is a third (this one is a Python shell, IIRC, not a library). I had some fun writing an experimental thing myself, related to this area (but not exactly the same), called pipe_controller:Swapping pipe components at runtime with pipe_controller:…Using PipeController to run a pipe incrementally:…>(C) For high level language, easy access to all of the old Windows System32 calls for writing old Windows graphical user interface applications?First, you don’t need to do it that way, unless you need high performance or low-level system access. Python has bindings to higher level (than the Win 32 SDK) GUI toolkits, including cross-platform ones, such as wxPython (a binding to wxWidgets), PyQT (a binding to Qt), recently Kivy (also on mobile), etc. Also, Python on Windows comes bundled with Tkinter, a binding to the Tk toolkit that comes with the Tcl language. It’s fairly easy to whip up quick GUI apps with the combo (Python and Tkinter), once you understand its GUI concepts like packing, grids and so on.Second, Python has good FFI (Foreign Function Interface) support (aka ability to interface with other languages, mainly C and C++, via a few ways, like the native Python to C API, ctypes, CFFI, SWIG, etc.), so it may be possible to wrap Win32 API calls, whether GUI or not. But it may be somewhat tedious and have some issues related to the GIL, GC, etc. I would stick to the higher level approach with Python (wxPy, PyQt), or, go to full lower level with Qt and C++ or wxWidgets and C++.>(D) For reading in data, say, comma separated values, name, type, value triples, XML, or something similar, drawing graphs and converting them to, say, GIF or PNG?Yes, Python has good support for data munging / transformation tasks involving various data / file formats, including CSV, XML (both SAX and DOM approaches), TSV, JSON, and others. In fact, my xtopdf toolkit does a lot of that sort of data conversion / reading stuff as part of the features it has of converting various formats to PDF. See:…and…Python also has many libraries for graphs and charts, multimedia, etc. E.g.:pngcanvas, a pure Python PNG library:…PyPNG, to read and write PNG files with Python:…pypng, pure Python module to encode/decode PNG:…You can also write your own custom code in Python to do that sort of stuff, using lower level primitives like graphics libraries to draw graphs and charts, if you want finer control or to do things your way.>Q 2. Since I’m programing on Windows, should I use Microsoft’s Iron Python which, IIRC, is compiled instead of interpretive, and has ready access to all of the .NET Framework software objects?Don’t know the answer to this one since I have not used IronPython. But I have heard that in general, it does support more .NET features than CPython.>Q 3. How would Python compare with Microsoft’s Power Shell in suitable uses, ease of use, features, capabilities (what it can do, not in the sense of security and access control lists).Can’t answer since I have not used PowerShell. What little I know or have read of PowerShell is that it is powerful (as a shell, maybe more than Unix shells, in fact I read some Unix experts were involved in designing it), don’t know if it is powerful as a general-purpose programming language, which Python is), so it may not be apples to apples comparison. Each may have their pros and cons.>Q 4. How good is Python for writing programs to make use of Microsoft’s SQL Server?I haven’t used Python with SQL Server, but it is likely to be okay-to-good for that, because both are very popular, so there is likely to be a driver. See below.Python’s database access is via the DB API, which is like that of Perl, JDBC and ODBC – a two-layer model with a driver manager and drivers for different DBs. You need a Python DB API driver for SQL Server.>Q 5. Suppose want a much better program than the old Linux Grep or the Windows Findstr. l would Python be an easier way to write such? E.g., for some collection of text files, find all files that have in one or more paragraphs at least three of the following list of 15 words?grep and findstr-like tools are quite doable in Python. In fact I recently did a simple fgrep in Python, may blog it after a bit. Quite small and simple code, and I have figured out (thought not implemented yet) that enhancing it for some more grep features would not be too hard. The custom find example you mention is also doable without a lot of code. That is one of Python’s highlights, conciseness and developer productivity, once you know it well, of course, to achieve which is non-trivial, because it is a deep language, having been around for, and evolved over, many years – over 20).>Q 6. Can Python get to some lower level stuff, say, on what processes are running, virtual memory paging rate, I/Os per process, how real main memory has been allocated, what the TCP/IP stack has been doing, e.g., time outs, errors?What process are running – see psutil. It can also tell you a lot of other stuff about your system.But in general, Python may not be too a good fit for this kind of work, AFAIK. You can definitely get many kinds of system / OS info with Python, easily, and even manipulate the system some (e.g. Python has support for some Unix system calls in its standard library, and in fact I have written many command-line utilities in Python, some of them related to OS manipulation), but digging deeper into the guts of system software like OSes and device drivers, is better left to lower-level languages like C, C++, Rust, etc. That said, the project Lawrence and I worked on (which he mentioned in this post) involved some TCP/IP packet processing stuff, which was done in Python.However, to get around this limitation, in general, we have the option (via multiple means) of writing the higher level parts of our apps in Python and linking them to the lower level or more performance-sensitive parts in faster languages like C, C++, Rust, etc.

          1. Vasudev Ram

            Also:>In fact, my xtopdf toolkit does a lot of that sort of data conversion / reading stuff as part of the features it has of converting various formats to PDF.See:…for my posts about xtopdf ,and also:…for my posts about Python (which are a superset of my xtopdf posts, since xtopdf is written in, and used from, Python)

          2. sigmaalgebra

            REALLY nice! Very good to have and to know. A keeper, read, indexed, etc.You confirmed the rumors I’d heard: Python comes with a lot of darned useful, amazing “packages”.I’m not much interested in writing graphical user interfaces: E.g., for my Web site startup, I’m, of course, writing server side code with little or no user interface. The only significant user interface is for the Web site users, that is, the Web pages, and those are just simple HTML with minimal graphics — the only graphics are a PNG file for the logo and whatever the ad networks send.But occasionally I want to produce a little in graphics. So, some nice, easy, capable ways to read in data, generate some graphs, and output GIF files, say, with more control than I get doing graphs in Excel, would be nice.I tried once in Windows using some Windows graphical things, something Microsoft calls GDI+, got some okay output, but kept running into problems, e.g., my code could not accurately determine the length on the graph of a character string. Sure, the lower level C, C++ calls to the system32 stuff would let me have as much control as the high end Windows graphical programs — Acrobat, Firefox, etc., but I don’t want to dig into that old stuff for just occasional side projects. So, if some parts of Python have made such graphical work easy, then, okay.My scripting language is still Cowlishaw’s Rexx, actually Open Object Rexx, available for download on Windows. As I understand PowerShell, it has full access to the .NET Framework, and for many things that would be more powerful than Rexx. But Rexx is elegant.For getting at low level system things on Windows, there are some amazing programs called System Internals and written by Mark Russinovich at Microsoft. His programs show that there are some amazing calls into Windows internals. IIRC for showing system data, Russinovich doesn’t make those calls clear, but they must exist. Also on Windows 10, the version of the old Windows program Task Manager seems to be open source so also illustrates that there are a lot of fairly well known calls into Windows for low level system information.Well, then, if such calls are ‘open’, maybe Python has a path to them.Such things are sidelines for me now, but when my first server gets busy, and, horrors, maybe crashes for no apparent reason, maybe I’ll wish I could look at what is going on in the TCP/IP stack, etc.Thanks! You just unlocked for me the first door into Python should I need to walk in for real!

          3. Vasudev Ram

            You’re welcome :)Yes. Mark Russinovich’s Windows system internals utilities seem to be quite good. I’ve only tried them out a little. I think he was an independent earlier, I seem to remember his site was called or some such name, and that was when I tried those utilities out a bit. He later joined MS, I think.Update: I just googled and found this page which gives more info:

          4. sigmaalgebra

            IIRC Mark was long with Microsoft and did his System Internals as a side project on his own. Eventually Microsoft had and/or let him make it a Microsoft download offering. I believe I’ve heard that now Mark is head of some big technical thingy at Microsoft, maybe their cloud server farm. He seems to be very well informed on Windows internals.One way and another, for their cloud hosting, their MSDN documentation site, and whatever else, apparently Microsoft has some darned well run, huge, high end server farms. No doubt they are using Windows Server. So, if eventually my startup needs a few million square feet of server farm space, then I would expect that with my running some millions of instances of Windows Server and SQL Server I could get some good hand holding from Microsoft on how to run a good server farm. On PCs, I’ve always been with DOS, OS/2, and Windows and have nearly never touched the Unix side. So, as I wrote the code for my startup, initially the easy decision was to stay with Microsoft. Given that, it is good to know that Microsoft should be able to provide my operating system, etc. software for nearly anything my company might need.

          5. Vasudev Ram

            > I believe I’ve heard that now Mark is head of some big technical thingy at Microsoft, maybe their cloud server farm.Yes, his Wikipedia page says he is the Azure CTO:

    2. Lawrence Brass

      I worked with Vasudev last year. We met here, right at the bar. Absolutely productive, delightful and professional experience. He introduced me to the python world fast and helped us to design and validate the framework.We are amigos now :)These days we are moving into production the system he helped to build, a financial transaction sniffer and network health monitor that feeds splunk dashboards.

      1. Vasudev Ram

        Thanks for the kind words, Lawrence, mi amigo. They are valued, and the feeling is, of course, mutual.

  4. Frank W. Miller

    Unbelievable data. Really valuable and interesting!

  5. mplsvbhvr

    Can’t say for sure if there’s any correlation, but I started learning Javascript because of Solidity and Python is next because of Viper…Completely anecdotal, but fun to see I’m thinking similar to others in the field even if I have no background at all in it.

  6. falicon

    In tangentially related news…check out (from the makers of crypto zombies and being touted as “It’s like Steemit and StackOverflow had a Blockchain Baby”).Just launched today. A lot of rough edges, but really interesting idea and head start.p.s. Still massive amount of room in this general space, so I don’t actually see it as competition for Stack (at least not any time soon).

  7. jason wright

    “White or of European descent” – what distinction is being made here?”Black or of African descent” – and also here?What does “Hispanic or Latino/ Latina” mean?

  8. sigmaalgebra

    For the geography, the circle for the US seems to be the largest, but the one for India seems to be the second largest. But for the ethnicity, India does not seem to be the second largest? Hmm ….

    1. Lawrence Brass

      I thought the same thing.I guess one have to apply the silicon valley bias correction factor. 🙂

  9. Matt Zagaja

    Always interesting to see how the developer technologies that excite me are different from most folks. I know VSCode is really cool but I cannot abide an Electron app.

  10. Rob Underwood

    Mark my words — in one year you’ll see ReasonML, a derivative of OCaml which has been developed primarily by the React team at Facebook and which uses an open source technology developed at Bloomberg called Bucklescript, on this list. ReasonML compiles down to ES5, bytecode, and native using much of the existing OCaml tool chain. It’s the new hotness.

  11. aminTorres

    https://uploads.disquscdn.c…You have no idea how happy this makes me. :)This means that about 27k developers visit the site. Even if that is wrong by 50% still makes me happy.

    1. PhilipSugar

      I love the Dominican Republic. I go twice a year. Such an easy non-stop from Philadelphia. Two a day.

        1. PhilipSugar

          Thanks so much. What a nice looking place. Booked at Punta Cana for the first week of April. Beaches and people there are so nice. I know it’s touristy, but I just love relaxing.

          1. aminTorres

            That’s cool, – the country can use all of the business it can get… 🙂 I advice people to avoid super touristic areas in general.Unless is something you have tried before and you were happy with… (which seems to be the case here) but unfortunately many people these busy touristic areas who provide services around the industry do not have the best practices. They think they’ve figure out the game and they assume/know most tourists they will likely never see again. So sadly, the hunger and cloud of judgement that comes with living in a economically and politically troubled place kicks in and people engage in abusive behaviors.I chatted with The GG a few years back and she told me she had a troubling experience at a place she rented in Punta Cana, I was sad to hear the details. but, to some extend this happens everywhere but it sucks more when it happens to people you know.Some Americans in the town I grew up built nice homes near the beaches and they keep a staff that is well trained over the years.I friend I went to high school with works at one of these homes for about 10 years now. They have lots of repeat customers – mostly families – which to me is a sign of good service. :)Would love to hear about your experience when you come back.Best wishes. If Punta Cana were closer to my town, I would have suggested you to visit a few people who could make you some delicious local food 🙂

          2. PhilipSugar

            I would love to get a place there. Been five times. The reason I like Punta Cana is that the flight is just so easy. When you have a wife and kids just getting on a plane and in 3.5hrs being where you need to be and then a 20 minute drive just really makes the vacation.The reason I do resorts: Going to a resort is a vacation for my wife. No cooking, no cleaning, no worrying about activities for kids. Yes I get taken advantage of somewhat. I know that. It used to bother me but it really is just a cost. I bring $300 in $1 bills and my family tips well. We are treated sooooo well. (Europeans hate us) We go to the same place, my name is easy. They call me big Poppi and the interaction with the staff actually makes part of the vacation. I get great food, awesome beaches, super relaxing.My final point is this: Go to Disney World. The amount you are going to get “hustled” is an order of magnitude more. The princess that fairy that pushes your kid to buy a $25 Chinese piece of shit wand and makes you feel like guilty if you don’t and your daughter just looks at everybody else and begs. The $20 hamburger that was sitting on a steam table? The crappy room that gets you a “pass” to the fast lane but costs $600?Hey you beat me for $10-$20 . Ok same about as Disney. But the money is going to support your family, and it is on a scale lower in sophistication. Sure that horseback riding tour is overpriced for my wife and daughter, yes the dune buggy tour is expensive for me and my son. Yes, you will get hustled a bit, but I have found the service is so good. Yes my son was crestfallen he was not old enough to drive his own dune buggy, but at the end of the ride the supervisor let him drive his around with him in the passenger seat (my son was overjoyed and that was before my tip)Love to chat. My name above at the gmail service.

  12. Lawrence Brass

    It is an excellent survey and report.I like the “wanted” lists and consider them as indicators or predictors for future growth.The following is the chart i liked most. It looks like the molecules of our digital world and gives a hint of the affinity that technologies have with each other. Sometimes by design and sometimes by chance. Technology is alive. https://uploads.disquscdn.c

  13. creative group

    CONTRIBUTORS:Any dedicated Daytrader will always be cautious of anyone having skin in the game who profusely pumps companies endlessly.We can’t fault Fred for his promotion because he provides a disclosure. The contributors who post this is the greatest ever survey/report shows their ambivalence and actually outlines how horrible the software industry lack of inclusiveness to minorities and women is actually shameful. #notincludedThe we should do better without doing better is misleading of words and actions to all decision makers. #notincludedThe usual suspects will defend the indefensible. #notincludedTo whom much is given much is expected. If a person has never been excluded they have no idea the feeling of those being excluded.Captain Obvious!#UnequivocallyUnapologeticallyIndependent

  14. Aria Zhang

    The survey obviously has not received much response from China.Therefore, the ethnic diversification of the result only reflects a picture of English-speaking world. However, considering China’s role in software programming, we should be cautious in interpreting the results.