I've seen so many posts like this talking about programming language demand as a number of positions available - but I don't think I've ever seen one talking about programming language supply. JavaScript is certainly in high demand, but the market is also brimming with JavaScript developers.
Some other languages on the other hand, are far less common, but because they occupy a specific niche, might actually be more "in demand" (ie, less competition for an applicant, higher salary).
That's an important distinction, but I think it's getting more common to post for positions you have filled so there are resumes available if someone takes flight. Resumes don't mean someone is available either.
Sure. Would you agree that this is something a JS-centric company is more likely to do (than a java one) because JS engineers are on average younger, and their life is more volatile, and also because the company itself is younger and more volatile?
Probably because it's difficult to determine the supply of developers for a language. Job boards tend to be companies posting jobs, not candidates posting (public) resumes.
Even if you had resume data, having a language listed as a proficiency doesn't mean that person is in the supply of developers for that language. Likewise, sometimes people get jobs for languages they don't know. I hadn't used golang before my current job, so I wouldn't have been in the "supply" of go developers despite being hired to use it.
The market may be brimming with js developers, but the bulk of them cant tell the difference between == and ===, cant get grasp the this/bind scope or lack basic understanding of inheritance, polymorphism and composition (the npm registry is full of copy pasted modules with small changes and a new name).
Getting js right is hard. Thus, competition for _skilled_ devs is rather low.
Same goes for php. Writing code full of sql injections, no design patterns, and no scalability is easy. Most php devs are limited to Laravel or Sympfony and have no clue about the language’s inner workings. Doing it right is hard and if you do you will get nice returns.
But i do agree about niche languages. The number of jobs available means nothing. Whats relevant is the (applicants/ number of jobs) formula.
J++ was an actual JVM implementation, c# ran on the .net runtime that was being developed at the same time for the next generation of Visual Basic. While the languages looked superficially similar, the runtime and base class library is significantly different.
After the sun lawsuit, Microsoft dropped j++ and released j#, which allowed for a weird hybrid of Java but with the .net class library running on the .net runtime. This was short lived because why use this over c#, and has been mostly forgotten to history.
Anders Hejlsberg worked at Borland for many years and was heavy into creating Delphi. He was hired away from Borland to Microsoft. Having said that, it's more the IDE than the actual language. Delphi and C# aren't similar syntactically. C# is a C styled language while Delphi is a Pascal styled language (well it is Pascal/Object Pascal in everything but name). There are some similarities though, for example, C# has standard data types along with objects like Delphi where Java does not (at least last time I checked).
Visual Studio early on and Visual Studio today are very similar to the Borland dev studio when Hejlsberg was working there. I've used both languages/IDEs in my career extensively, albeit I haven't used Delphi in probably 15-20 years.
"Job offers with fuzzy language requirements were discarded."
Doesn't this mean that they're only counting cookie cutter ticket grinding dev jobs? Which will be heavily biased to the current top ~3 corporate use languages.
IME in good dev cultures the teams have a lot of freedom in picking the tools, and people are recruited assuming they can pick up new programming languages on the job.
I am suspicious of that mirrored graph between Java and JavaScript. I see no real world reason for they behave in such symmetrical and substituting way month over month. Are you sure there is not a parsing error when analyzing those two languages?
A lot of HR people still don't know the difference in 2022, it's quite common to see job ads listed as Java/Javascript, then you only find out when talking to them they specifically mean one or the other.
Many job ads seem to do things like confuse C and C++ (and then you read the ad further and find its a C# project), or just slap a long list of technologies, I won't be surprised if these ads confuse Java and JavaScript too and list them together.
Since those are percentages it could simply mean that a great amount of Java positions opened up in that period, stealing percent points from other categories. This explains the symmetry, but not the causes.
May it be a huge amount of positions have been incorrectly labeled for Java instead of JavaScript (you know… Ham / Hamster) and then corrected a month later?
Or the absolute number of Java listings is pretty much constant (permanent, big corps that just leave listings online all the time), and as smaller web companies add JS listings, and remove JS listings, that fluctuation makes the Java percentage move in inversion to the more volatile JS listings.
I read this as javascript possibly being the language of recent growth: It's relatively young, very much en vogue, and so new startups are likely to use it disproportionally (because the guys who have been around for a while will not just switch their millions of lines of code to js because it's now become a thing).
But startups are also more likely to falter quickly in a recession than the old guard. And that's where java (or c#) is overrepresented.
I don't get Kotlin's position in rankings like this one. I've heard people say, "oh, for new Android apps, you definitely want to use Kotlin." And (I asssume?) people are writing lots of new Android apps.
But this methodology has Kotlin at 0.35% of jobs! And for that matter, Swift is not much more, at 0.38%. Why are these flagship languages for large ecosystems (Android/iOS) not more in-demand?
I do Xamarin (Microsoft's iOS / Android framework) and the listings often throw in Swift / Java / Kotlin. Sometime's they'll throw in React Native and Flutter even if they don't use them. The Xamarin world isn't huge and I guess they just want to try to grab the attention of as many hybrid mobile devs as possible.
I think part of it is that there are a lot more old Android and iOS apps out there that still use Java/Objective-C. These apps are probably worth maintaining and updating, but not necessarily worth rewriting. (And, you'd still need to know the old language in order to do a decent rewrite.)
For iOS where I work, we use primarily Objective-C, with only a tiny bit of Swift. We actually have more Rust code than we do Swift. (The Rust code is for our cross-platform logic that's shared between Android and iOS.)
I think the Android side is all Java with no Kotlin at all.
I recently suggested to one of my consultancy customers (valued at $500mil) that they should hire Kotlin and Swift developers to improve the quality of their React Native app.
The response I got is: "nope, we're fine employing one mid-level React dev to maintain our app in both platforms."
They don't consider mobile apps being essential for their fintech business in 2022. I'm speechless but this serves as a hint as to why we don't see more Kotlin and Swift jobs: companies are cutting corners with React.
As someone who does cross-platform frequently now, after doing native apps for a long time, that gives me anxiety for that dev lol. I can't imagine how helpless they feel to fix the 1000 random issues they run against in cross-platform land.
Its the nature of the beast, as well as the market. The web market is massive compared to native app dev and app dev teams are minuscule compared to web teams. It would look less weird if the JS/TS line was actually split into the component frameworks (i.e. Vue/Angular/React/etc).
The key thing to remember in any "x is dead" meme is that all it means is "this isn't the new and shiny anymore".
Ruby is as good as it ever was, it's just not shiny anymore.
And god, I remember the shiny. I remember when Ruby on Rails came in the scene, and the whole fuss about Basecamp and how influential 37Signals seemed to be back then. Time flies.
Everytime I try out a new framework, I end up just switching back to Rails because "wow, I could do this all so much quicker in Rails...". Part of that is just a decade of familiarity, and part of that is the maturity of the framework.
The only thing that I've found to be more productive are the integrated GraphQL platforms (Hasura, Postgraphile, etc). However, these seem like they struggle to scale with team size.
Based on what I see around me Ruby is used by a couple of big companies, while something like PHP(another language that people generally consider as "dying", probably way more so than Ruby is) is used by a lot small companies. Would be interesting to see OP's data, but instead of amount of jobs it's amount of unique companies that offer these jobs to see if I'm correct in that assumption.
Eg. Go (say) is often a secondary skill listed with the real target, like Java. This means it's not as in demand as might appear on the surface let alone considerations of whether it is actually required by the recruiter or is just being used as bait.
That's concerning. I am a Javascript/Typescript backend dev trying to get a full-time Golang based role. I wonder if should look for a Javascript role which has some Golang on the side.
"DevJobsScanner has only picked the job offers that explicitly required a programming language. Job offers with fuzzy language requirements were discarded."
There's a handful of token jobs but even if I were interested in that, the pickings are slim. It's definitely not "no jobs" but there's a wide spectrum.
From my experience a lot of rust development is done in places where rust isn't really what they hire for. It's existing devs who start writing some greenfields projects
Elixir, Clojure, Rust. These were all great languages that were championed but nobody is learning them because there are no translatable jobs that you would find with ease in other languages like Javascript or Python.
The compensation isn't there to match it. Neither is it like a COBOL type of situation where someone's mainframe is running on super niche language.
Really? Speaking from the Elixir perspective, I get recruiters messaging me with surprising frequency about senior Elixir SWE roles with base salary up to USD $200k; it doesn't seem like compensation is the issue, at least from my experience.
Well, it's difficult to verify; as a survey participant you can enter whatever figure you want; and even if it were true, the sample size is very small to conclude anything reliable. Measuring the number of jobs advertised is more objective; and with a thousand or more jobs, the noise component should also be sufficiently low.
I think the real problem is not this. But that comparing these average salaries is meaningless even if the responses are truthful, the sample of respondents is representative, and the sample is large.
Clojure developers have a different distribution of unobserved skills and motivations than JavaScript developers or Go developers. So the difference in average salaries could be due to the robustness of the language, due to the pockets of the industry that use them, due to unobserved skills and related reasons, or even due to differences in how much Clojure vs. other developers value amenities like the novelty of the task that they are working on.
It would be interesting to see more than just a comparison of averages. But collecting good data to capture the finer details is hard.
How do you know the sample is representative? If I look at the data posted by the OP there are about 450 clojure jobs a year; if we assume that developers switch jobs all three years in median we get about 1350 clojure jobs in total; if we assume a very good response rate for the survey, about 3% of these developers or about 40 responded; this is even less than the number of states in the US.
Whether a sample is representative is a different issue from whether the response rate is high.
For example, you can have a population of 10,000 jobs, 9,000 of which is hiring for Clojure and 1,000 of which is hiring for Forth. If you sample the 9,000 Clojure jobs, then you might conclude that 100% of all 10,000 jobs are for Clojure. But in reality, only 90% are.
Instead, you can sample 100 of the 10,000 jobs at random. The expected value of the average of whether a sampled job is a Clojure job will be 90%. There will be noise but that can be statistically accounted for.
If the population that you want to draw conclusions about is, say, the complete universe of jobs ever offered in the US in 2021, it will be difficult to find either a data set that contains this universe or a data set that is arguably a random subset of the universe. So representativeness is hard.
You could adjust your population definition to achieve plausible representativeness. For example, take the population of all developer jobs at companies that had an IPO between 2018 and 2021. Maybe you have a way to compile this data set from some source. Then you limit the scope of your claims but you will be more credible.
Another thing that you can do is take an existing data set that you know to be representative and compare the distribution of job characteristics in your sample to that. For example, you might find that your sample is more likely to include web development jobs than your reference data set. Then you know that your sample is not representative, and you know in what way it isn't. Or you might find that your sample is comparable to your reference data set. This can give you some confidence that your findings generalize.
Agree; even if we don't know the distribution or representativeness of the samples we can do some guesstimates, as I just did; it's as good or as bad as what you see on Tiobe or other ranking services; the error is likely to be larger the smaller the population under consideration, which is why I'm rather skeptical about the salary statistics on Clojure.
I made two simplifications in my thinking. First, I assume that the number of Clojure jobs is constant over the average job retention period. Second, I assume that each job change is advertised exactly once, or the OP evaluation ignores redundant advertisements for the same job. If the latter assumption were not true, the number of actual jobs would be smaller by the redundancy factor. There are studies on the job retention period (e.g. https://hackerlife.co/blog/san-francisco-large-corporation-e...) from which I just took the longest one for a conservative estimate.
The number of responses to the Clojure survey is barely representative for the size of the job market; otherwise you would have to expect millions of respondents for e.g. the JavaScript survey; people responding to such surveys are likely fans of the technology, not necessarily professionals who were hired to use the technology; of course there are also professionals among the respondents, but we don't know the exact proportion.
It's been in the same position in previous years. In any case that's the best data I know relating to the claim about the compensation I responded to, if anyone else has better evidence to either direction, it would be interesting to hear.
Conflating JavaScript and TypeScript feels like a stretch.
On the surface, that appears to take virtually every company doing a front end web app and merge it with everyone doing JS or TS on the backend, when the rest of the languages are just backend.
The author doesn't seem to clearly list a methodology so I'm guessing a bit but...
I put a lot more stock in Stack Overflow's survey results, although I also wish they distinguished frontend JS vs backend JS usage in comparison to the rest of the languages which are only backend.
From what I've seen (I work in the field), surface level AI/ML skills don't really command a salary premium these days. So many people hopped on that train that the market is saturated.
I bet there’s a lot of weird effects in compensation. Probably dead languages with large, critical code bases pay a fortune, then there’s odd situations like ocaml that’s only used by one very successful organization.
Companies are notoriously bad at listing what they actually hire for in a JD. So unfortunately this might be misleading... It really only answers the question -- "If I want a job right now, and only for now, and I do not know any programming languages, which should I learn?"
It doesnt even give job stability because some of the highest languages may be on their way out, whereas the lowest might be just coming to popularity (and will rise for +10 yrs) ...
True that. One particularly odd situation is common in data engineering circles. Most DE descriptions require something like "experience with Java/Scala/Python". Notice that one of them is from a completely different family than the others. Very frequently nowadays they actually have a predominantly Python code base.
Searching by simply "Java" or "Scala" is nonsensical for exactly this reason. They are mentioned in a large number of job ads just in case. Even if they are not actually used.
Where did the scanner find 4,000 Rust jobs? During my recent job search, I found only five: zed.dev code editor, systeminit.com devops tooling, fly.io, Microsoft security tooling, and Tesla car simulation. I excluded cryptocurrency, embedded, and some really cool jobs in Europe.
This is only showing demand. It doesn't include how many people are already employed using these languages. Perhaps these languages are popular and self-taught among already employed programmers so demand is reduced.
How many of you are getting offers for a job where you don't know the language being used AND have less than 5 YOE?
My take on this type of study is that it's extremely noisy. There are tons of job ads that list 3-5 langs. These are basically recruiter dragnets and there are many of these postings out there. Of course JS would be listed in most.
I think this data would be more useful if we can filter by lang count in the job description. 1-2 langs I get, but more than 3 would be a red flag. It could be causing heavy bias.
This reminds me that C/C++ are probably the most successful programming languages in history. They are ancient compared with the competition but new projects are started every day in C/C++ and most of the worlds software runs on top of C/C++. JS/TS/Python for example runs on top of runtimes written in C/C++, running on top of an OS written in C/C++, running on top of a CPU designed with tools written in C/C++ etc.
some of this is as a result of a self feeding endless loop i.e in regards to JS.
using JS or more apt us JS dev requires / create complexity, which in turn creates more demand for JS devs.
But on a more first principles Basis - I have asked this question before [0]: Do we need web browsers ?
The web browser is the least closed most open distribution for software services ever. Anyone anywhere can publish and billions can access it without numerous gate keepers in the way.
It isn't the only solution out there. You can publish to the various app stores and abide by all their rules and regulations and upkeep requirements.
You can be confined to Linux desktop.
You can still get to the desktop via your own distribution. You can also side load Android and avoid the store.
It's all about observation bias. The SO survey wasn't representative in any conceivable way. Only active users (as already pointed out, the vast majority are passive users) who are passionate about their programming language of choice were sampled.
Most professional programmers aren't passionate about the language they use - it's a tool or a management decision, nothing to get emotional about.
The more "esoteric" or "exotic" the programming language, the more passionate its user base. Mainstream is defined by mediocracy and the average developer simply doesn't care.
In my experience, the majority of developers don't even log into StackOverflow, they purely use it as a read-only resource. I'd say even less actually fill in the survey!
Some of the most loved languages like Elixir, Clojure and Rust are fairly modern and powerful and fun to use. But they are also scary. Elixir and Clojure are functional languages, Elixir runs on a niche runtime, Clojure is a Lisp and looks foreign to people who are used to more common syntax, Rust has functional concepts baked in and is known to be hard to learn initially.
Unfamiliar + hard to learn = scary. Which leads to less overall popularity when it comes to actual jobs.
Wasn’t Rust the #1 on that poll? Yet the amount of code in Rust is practically none, and there are vanishingly few qualified/experienced Rust programmers. There’s a massive gulf between what language has the share of memes and which are used to write real code in industry.
My experience at a few places: devs try it out for hobby projects. Maybe get a POC in Elixir done at work. Ultimately the business continues to build on the existing stack. Hence, few job postings for Elixir.
I'm curious to know how much bias there is against smaller languages.
From my experience reading job postings, a lot of job postings for jobs where you may work with more "niche" technologies list more common ones beside what they are looking for and expect you to learn the technology during onboarding.
The whole idea of next.js is that it‘s the solution for backends and frontends. Using it for your frontend only makes absolutely no sense at all. Or is there something I‘m not seeing here?
You can use it as a standalone backend or you can just use it as a router / renderer "web backend" but then still have an API off somewhere else. That API could be a monolith, a bunch of microservices, power multiple front ends, be a legacy system etc etc.
When I said backend above I meant company backend / API / services.
Ignoring the API routes in Next.js is perfectly valid, the build system and built-in routing and utilities make it a better frontend-only framework than anything else in the space by far (CRA, etc.)
Using it with a separate backend/API is also sometimes necessary, difficult to use serverless functions if you want a WebSocket server or similiar.
Really cool to see, I’ve been exploring some scraping options in 2022 for something similar to this but in another niche. Would love to hear any broad details about what you’re using for scraping and data extraction.
Interesting evaluation, thanks. Are these only US jobs, or worldwide?
It's surprising that there is so much need for Ruby (even more than Golang) and Lua, and that there is five times less need for Dart than for Lua developers.
so if you know python/js you should be open to 50% of jobs. but that's without considering all the frameworks and libraries that job ads like to pile on to fill out that word count.
It might not be a cool language, but I'm surprised not to see VB anywhere here. It's still used in plenty of places, why does it not show in job listings?
The author talks about 7M jobs, but the graph of the most popular programming languages only lists around 1~1.5M jobs. What were the technical requirements of the other 5.5M jobs.
it used to be that Ruby was more popular but seems like Python ultimately eclipsed it. One thing that really hurt Ruby was there was just way too many ways to accomplish the same thing and there were lot of opinions and ego around which was better where as Python has established standards like PEP8 that champions readability above all
There are definitely as many ways to do things in Python if not more... I've worked extensively in both languages. Frankly, the Python ecosystem is a lot more complex to navigate (dependency management is a nightmare) and much less friendly.
I think the biggest difference is (1) Python more institutional backing (2) Python is taught in academic settings (3) Python has more robust scientific computing (4) Node stole Ruby's thunder in the web application space (5) Python works out of the box on Windows.
I don't know if it's better these days, but dependency management was one of the reasons I stuck with Ruby over Python back in the early 2010's. I never had issues with it in Ruby, but was constantly battling with dependency management in Python.
I think #5 is huge, until Windows Subsystem for Linux came out there was no easy way to compile all needed gems on Windows. Most Ruby/Rails tutorials assumed you were running *nix and DHH's dismissive attitude to Windows didn't help. Git Bash for Windows bridged some of the gap and probably Cygwin did the rest, but batteries were not included for Ruby on the most popular OS.
One thing that really hurt Ruby was there was just way too many ways to accomplish the same thing
I dunno, to me it feels like that's why Python won. Python is more pragmatic, flexible, less wedded to ideals, and, I think, these things allowed it to pick up a lot of entry level developer attention especially in areas like data journalism, data science, ML, etc. which stole a lot of Ruby's "good for beginners" thunder. This is now paying dividends as all those new folks have cemented Python's now bread and butter projects like PyTorch as must-have skills. Ruby not ever really caring about Windows didn't help either.
From my experience, it's mostly older people (50+). But my experience is also that companies still using Fortran understand that they need to train new hires to use it. So the "qualifications" for these jobs is probably more about background than programming language proficiency.
15 years ago doing JavaScript or Python you’d be classed as a low tier web developer on the bottom of the pay scale.
Searching the job market the past year everything is JavaScript or Python because of the low barrier of entry. With the demand to hire high numbers of people and lack of people, these roles are now at the top of the pay scale.
Maybe it’s AU specific but if I was to take some specialised role requiring deep knowledge and experience, not using JavaScript or Python, I’d be on at least 20% less than a role essentially knocking out a ReactJS frontend or NodeJS crud services.
It’s dilemma I had the last year. Recruiters phoning with somewhat highly specialised roles that interests me, talent pool is extremely small for the roles, only to be told that the pay is at minimum 20% less than what’s available for nondiscript JavaScript/Python roles. Unfortunately for me I need to enjoy work or suffer mentally, and while I have used JavaScript, Python, insert name of frontend/backend framework in either language, extensively, it feels like torture having used better and knowing I can do better.
How so? The HN title, "I scraped 7M dev jobs for 8 months and these are the most demanded languages", is different from the article title, "Top 8 Most Demanded Programming Languages in 2022", but they're quite similar and both describe the article content.
Just yesterday someone made a comment about how someone could spend years laboring away on some super cool piece of tech and someone on HN will ignore it all just to point out how the kerning is wonky.
Your comment is exactly what they were referring to. No comment on the project, the data, the conclusion. No related tangent, personal anecdote, or anything else that would generate conversation. Instead, hyper focus on finding some minor flaw that has absolutely 0 impact on the substance or delivery of the project and be sure to call it out in a condescending way.
Should Javascript be written as JavaScript? Sure. Does it materially change anything at all about what is being presented? Absolutely not.
FWIW, my comment was made in good faith to help the author create a better impression with folks who are steeped in those two technologies. At the same time, I can see how it can feel pedantic.
It's stylized as such - you are not forced to type Yahoo! every time you mention the company just because they want you to. Same with GNOME and Gnome, etc.
I was lucky early on in my career that I started with JS. Used to write code strictly for IE6 then moved to ‘cross browser’ compatibility code.
These days we’ve come so far and you can reliably write JS that works everywhere, so no need to cater for niche browsers and target a specific browser.
I dabbled in PHP but stopped using it since it’s hard to write secure code with PHP (although you can audit it).
Settled with Perl and Python for small hacks that automate mundane tasks, but am far from creating anything elaborate with them. My Python scripts are rather small. One thing well and all that.
Yet I often come across websites that only work in Chrome. I think the "no need to cater for niche browsers and target a specific browser" isn't as true as web programmers would like to believe.
Some other languages on the other hand, are far less common, but because they occupy a specific niche, might actually be more "in demand" (ie, less competition for an applicant, higher salary).