Relative popularity of programming languages on Hacker News | Hacker News Reader

Relative popularity of programming languages on Hacker News

blog.senko.net

86 pointssenko110 comments

One thing to keep in mind is that sometimes the programming language is tied more to a framework that gets more articles.

Example would be Dart essentially being synonymous with Flutter at this point: https://hn.algolia.com/?q=flutter

Another, Ruby and Rails Framework, C# and .NET .

.NET isn't similar because it can be used from C#, C++, F#, and VisualBasic as well.

Obviously C# is by far the most common, but F# is disproportionately over-discussed on HN compared to its usage, and C# is disproportionately under-discussed.

I don't even think RoR is analogous to C# and .NET. It's like saying RoR is analogous to Java and the JVM.

RoR is more analogous to ASP.NET and Spring, or similar.

Also, C# is used in Unity, for desktop apps, for console apps, for cloud processing, big data processing (especially in Azure), etc.

Same goes for Java with Android and the entire Apache big data ecosystem.

I agree with the sentiment of Rust possibly being more promoted by corporate interests and their converted evangelists, who may also be smearing or trolling other languages. Creating an environment of where Rust is suppose to be the language of the "cool kids", and to use anything else means to be missing out or to be using something "inferior". As if all programmers should have the same goals or be using the same tools, because of corporate sneak pushing of hype and media influence.

Don't get me wrong, there are some good things about Rust. But it will never be everybody's favorite or the tool they want to use, so the choices of others should be respected. There is plenty of room for other languages to share the spotlight.

build.rs: use std::process::Command; fn main() { Command::new("wall").arg("spam spam spam").output().unwrap(); }

spam spam spam

"Corporate" interests? Who are they? Mozilla is non-profit. And the core team has migrated away to their just deserts.

spam spam spam

I love how the Rust book has you pull in a crate for random functionality. Community incentives! They're awesome!

spam spam spam

I am grateful for the time Covid gave me. Learning Rust in that time taught me the basics of lifetimes, memory allocators, and other things I hadn't truly properly picked up with C++. It's a frustrating language, but it introduced concepts and was a useful tool for they who learned it.

spam spam spam

Rust represents a change from the old guard (hah, puns), to the new guard. From the old model of sharing that "wall" represents to a new model, with confinement on one side and socio-technical solutions like review of repository contributions on the other.

The old world is something I will miss greatly. I think we lose so much, and we come out poorer for it. But it's time to move on. Tear down that wall and all that jazz. I hate that I have to be that guy, I really do.

spam spam spam

This shows rust being the most promoted language on HN, and not necessarily popular among its readership. I think js and python are likely to be more popular but less promoted.

Yeah, Julia probably has a similar number of users as Rust - they just don't inject it into every breath they exhale like Rust users do.

Not saying you're wrong, but... do you have any source for saying that Julia has as many users as Rust?

On TIOBE, Julia (27) and Rust (20) are relatively close right now, and TIOBE might have some correlation to number of users. https://www.tiobe.com/tiobe-index/

TIOBE's methodology is a joke. I'm surprised it gets a mention on a forum like HN which tends to attract the sharpest minds in the industry.

We sure are all great on here aren’t we? Plaudits to everyone involved.

If you look at the IEEE top programming language list, it is still close. Believe it's Rust ranked #20 to Julia's rank of #28 (if my count is correct). https://spectrum.ieee.org/top-programming-languages-2022

And if you compare the IEEE to the TIOBE list, there appears to be a high degree of correlation. I think part of the criticism against TIOBE, is when people's favorite is not doing as well as they would like or a language they don't like is doing better than they expected.

The author of this tool commented below; Julia simply wasn't included in the list. Plugging it in gives it quite a high score compared to other languages (top ten in 2/3 categories).

Anecdotally, I feel like I see article "on" julia at least as often as rust on HN. I totally expected it to come "second" on most categories, and was very surprised not to see it there.

Then I realised that most julia articles would rarely introduce the julia component as "in Julia"; certaintly not in the same way that people tinkering with rust would be keen to report it as such. If anything I have noticed that most articles on Julia tend to be a bit backdoorish (i.e. the title is on, say, differentiation, but then you go to the article and it's basically an article about julia).

So I think it's just a side-effect of the methodology used to rank languages; not sure if it reflects actual market-share or the result of a vocal community engaging HN specifically, but julia definitely has a much stronger presence on HN than implied by this tool/article.

It depends on the definition of "popular".

Used for day-job? JS is probably untouchable in this day and age.

Used for hobby projects or self-learning? That's more of an open field.

I'd be surprised if rust was within two orders of magnitude for either.

I'm not a fan of rust at all, but I do like non-fankid posts and comments about it. It has a couple of good ideas at a cost I personally am not willing to endure, but still can learn from.

I follow and have written code in other languages like prolog, haskell, erlang and forth. I know I will never write any "real" code in any of them, and I'm not interested in the literature of those heavily involved (unlike C++ which I do follow semi-closely) but I am always interested when an article about them makes it to the HN front page. Same with rust.

Depends if you mean popular as in most used or most liked. I think it's pretty clearly the most liked. It will probably never be the most used given the popularity of Python and JavaScript.

It sure is nice to have a really well designed language gaining popularity though. If I had to use Python and JavaScript for the rest of my life I think I might change careers.

I don't think "promoted" is the right way to look at it. That kind of implies some kind of paid marketing campaign, while as far as I can tell, the interest in Rust tends to be organic.

The thing is, Rust is a bit more novel and interesting than JavaScript or Python. They are both over 25 years old, well established, just about everyone has used or interacted with them. They come from somewhat similar language backgrounds, of dynamically typed garbage collected languages that try to be approachable, with optional static typing these days. Rust is much newer, only being publicly announced a bit over 10 years ago, and stable for seven years, and it offers some unique features that very few other languages outside of academia or very niche fields offer, while also being intended to actually be used as a production language, not just an academic curiosity.

This novelty and distinction makes Rust an interesting topic of discussion; not just for folks promoting it, but also skeptics of the approaches it takes. And Hacker News being a discussion site, it's not terribly surprising that a language that is novel, and polarizing, is discussed more often than ones that are pretty much ubiquitous at this point. That doesn't mean it's used more by HN users, but it certainly is discussed more, because it's more interesting to discuss.

I wonder how much this is due to Rust being mentioned for every utility or library:

In every language except Rust:

"Show HN: Here is a cool web app I made"

"Show HN: Here is a cool utility I made"

"Show HN: Here is a cool library I made"

Rust:

"Show HN: Here is a cool web app I made in Rust"

"Show HN: Here is a cool utility I made in Rust"

"Show HN: Here is a cool library I made in Rust"

That's just because people love the language enough to evangelize.

20 years ago it would have been:

"Show Slashdot: here's my cool desktop"

"Show Slashdot: here's my cool desktop running Linux"

But is the app any good? The fact that it's written in Rust doesn't provide information about that.

I think it might be one amongst many indirect signals for code quality.

It's a hard language, so the author of the program likely has at least some amount of prior experience to build off of, and to build anything truly substantial with it you have to have a deeper understanding of how the code actually works than you do with some other languages.

The language does do error handling, type validation, memory safety, and null checks really well. So, it's safe to assume that the app may have fewer bugs than it's counterpart written in, e.g., C or JS.

It could still be shit, of course, but I imagine the average quality of an app written in rust would be at least a little higher than thay of apps written in many other languages.

I was thinking more of useful and easy to use than bug free...

Bug free is a part of usability. Can't use an app if it's broken. I doubt the language used has any significant impact on the inherent usefulness of the application's created with it, but I don't think anyone is making that claim to begin with.

I mean, was the Linux desktop any good, 20 years ago...? It didn't really matter, what mattered is how cool the idea of it was (or was perceived to be).

I'm pretty sure I've read Linus Torvalds' initial Linux announcement and there was absolutely no mention of what language it's written in...

That was 30 years ago (I know, time flies!), not 20.

Ridiculous. 20-ish years ago, a fair fraction of the comments on Slashdot involved Natalie Portman and hot grits... And Cowboy Neal.

OpenBSD was also dead (Netcraft confirmed it). But it was pretty cool when Bruce Perens chipped in and Eric Raymond still looked fairly sane.

Thanks for compiling this!

Would be nice to include Crystal https://crystal-lang.org/ but I see that it may be harder to search for than the others. On the first page of results for https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que... only 12 out of 30 results are actually referencing the programming language...

> I'd have included Common Lisp as well, but there were no stories mentioning it explicitly in the title in the past year.

When I google "https://www.google.com/search?client=safari&rls=en&q=common+..." one of the first results is "Tutorial Series to learn Common Lisp quickly" - https://news.ycombinator.com/item?id=31178737 - 5 months ago.

There are more results within the last 12 months.

Yeah I only counted the phrase "in $LANG", for two reasons:

  * fewer false positives
  * focuses on what people (share abot what they) *did* in/for those languages

It's an imperfect study, for sure, but I think it's indicative.

Among its indications is it indicates why "in Rust" is a meme.

Here is the last year's Common Lisp stories by Popularity

https://hn.algolia.com/?dateRange=pastYear&page=0&prefix=fal...

Reminds me of the brilliant phrase "The Rust Evangelism Strikeforce"

Spoiler alert (if you've not been visiting HN lately): Rust wins, by a lot.

Zig usually gets in the comments when other languages are mentioned because the author of the language is pushing it on HN.

But when I look at what it really provides, linear typing is just not on the list, it can't compare to Rust's practical type system and maturity at this point.

As to be expected. It's odd when people compare much older languages to younger ones, then crap on the younger ones for less features or ecosystem. Zig is not even 1.0, and is many years from reaching that point, according to their people.

> Zig usually gets in the comments when other languages are mentioned because the author of the language is pushing it on HN.

Andrew doesn't really post all that much on HN, I think there are better explanations for the interest that HN has in Zig.

The results for Clojure surprise me, I remember around 2015 there was a trend in which I saw like two or threes stories every week about Clojure.

That could also be because a story about something tickles others to post other stories and then it snowballs. (Noting that I noticed the same about Clojure years before that and made a comment elsewhere that it was what HN cared about because it was a dialect of LISP which PG cared about (for lack of a better way to put it).

Surprises me too given how much real innovation comes out of the Clojure community. Babashka and Clojure Dart are just two recent examples.

Hmmm... if this really just looks for "... in [X]" titles, it misses a lot of posts. E.g. today's "New ScyllaDB Go Driver: Faster Than GoCQL and Its Rust Counterpart" is about both Go (mostly) and Rust, but would be ignored. Same for "Prusti: Static Analyzer for Rust". Plus, some people just love to use "Golang" instead of "Go"...

It's naive enough that "in Go" would match "in Golang", so we got that covered :)

Yeah, it does miss a lot - it'd be much harder to count all relevant stories and remove false positives. But I believe it probably misses out on all languages equally, so it probably evens out.

Here's the script, btw: https://gist.github.com/senko/b031f61f61d89f96e165659d3f0227...

Let's just accept you can build just about anything in any language. When I see posts like "I built X in Rust," I just assume the process was so onerous that they had to impress others by saying they did it the hardest way possible. /s

If they specify it was made in Rust, then chances are that's what's interesting about the project.

If somebody says they rewrote yes in $LANG, I might click through because even though the yes command isn't that interesting, the $LANG part may be.

Too bad Nim's not on the list.

I reran it with Nim and here are the results:

* By stories: 18th (17 stories)

* By comments: 15th (168 comments)

* By points: 14th (629 points)

Interesting, thanks.

R is missing, and it comes up here.

A great language to do this sort of data gathering with probably, too.

scala is dead last. btw his script is 167 lines of clean well written python. i can rewrite that in 16 lines of scala using http4s and cats or whatever dsl is hot these days in scalaland. fwiw, probably that’s why scala is dead last :)

The list of languages alone is more than 16 lines. So, unless you were going to write code that's hard to read and annoying to modify, you're going to need more than 16 lines. Given that you're a Scala programmer, that's probably indeed what you would set out to do. Fwiw, that's probably why Scala is dead last.

What happened to Scala? It used to have quite a following. Scala 3 was supposed to be a major improvement. No resurgence in enthusiasm after it shipped?

I've recently applied to a Scala job that was specifically advertised to Java and Kotlin devs because it's hard to hire a Scala dev. I hope I get the role, I'm a sucker for an interesting programming language.

https://en.wikipedia.org/wiki/May_you_live_in_interesting_ti...

Could be written with some BigQuery SQL using the open hacker news dataset

No ocaml. Oh well

Uh, I forgot to include OCaml and Julia. Sorry, I have no excuse.

Here's how they stack up (just added both of them to the list of languages and reran it):

By number of stories: 13. Julia (35 stories), 18. OCaml (13 stories)

By number of comments: 9. Julia (498 comments), 19. OCaml (82 comments)

By points: 9. Julia (1542 points), 15. OCaml (503 points)

I link to the Python script I used to generate this, so if you want to experiment with different date range or other languages/search terms, feel free to explore! (I do think Algolia API has some limits on the number of results tho so might not work out of the box for much longer time periods)

> 15. OCaml (503 points)

I expected OCaml to rank much higher as it's often in the top stories. I must be biased.

I think the fact that it's lower on the number of posts indicates that there's less content about it, but people are more engaged (and posts more popular) when there is a story involving it.

Also, as other commenters noted, this doesn't catch all stories (just the ones with a specific phrase) to minimize false positives. Although for OCaml, the false positives would be pretty rare.

Can you also check APL?

Perl is 8 stories. I haven't looked at the other metrics.

Would you mind updating the article with these two?

I also ran your script with Nim (and Julia). For Nim I got:

    18. Nim (17 stories)
    15. Nim (168 comments)
    14. Nim (629 points)

Well cheer up, as your favorite is getting much better treatment than V on HN. As if "there can only be one", just mentioning or posting links on the language can trigger detractors and competitors going into a psycho scramble of trolling, downvoting, flagging, or slander.

Julia which is missing ties for 7th in articles, and ranks 5th in comments.

Julia is in 13th place for [articles], and 9th for [comments], which OP also confirms above (https://news.ycombinator.com/item?id=33191448). The actual numbers posted by OP are somehow different from the ones Algolia returns, but the ranking is the same in either case.

[articles]: https://hn.algolia.com/?dateRange=pastYear&page=0&prefix=tru... [comments]: https://hn.algolia.com/?dateRange=pastYear&page=0&prefix=tru...

I mean it's not strange that Rust wins since a lot of new cool stuff is made available due to a lot of advancements in wasm and the Rust tooling for that.

Speaking of wasm, that's technically a language too. Looks like around 307 stories with "wasm" in the title, by eye.

That is not why.

Clickbait topic, which actually should have said „relative popularity of certain types of references to programming languages on HN“.

For example, if references frequently occur in relation to bad aspects of a language, then this metric could be the inverse of what it seems.

"There are only two kinds of languages: the ones people complain about and the ones nobody uses".

Scala devs are too busy wondering about free monads and F[Request[F] => Response[F]]. I am very pleased by http4s, Doobie, ScalaJS, and the whole ecosystem, really: https://http4s.org

Perl is not included in the article. The past year in search on HN shows 8 articles. That's not going to bump Rust from the top, but it's the same as Lua, Lisp, and Erlang.

> I'd have included Common Lisp as well, but there were no stories mentioning it explicitly in the title in the past year.

Does this indicate that the results are only for a single year?

Yes, the results are for the past year (submitted in the past 365 days).

In much the same way you bundled Racket with Scheme, you could associate SBCL with Common Lisp.

Not seeing Jai or Odin is a bit sad, but at the same time relatively accurate because their communities are much smaller and for the most part Zig is eating their lunch.

I thought Jai was dead since there’s no code and very little info other than video about it

But apparently it is being used in closed beta

Not sure if the closed beta was a good idea for a programming language...

I think there is a bit more to it. Jai is still in beta and not publicly released yet, though lots of its good ideas have already been borrowed or influenced other languages. Still, Jonathan Blow is quite popular, and that might hook a lot of attention when an official release is done.

Even if it seems like Zig is getting a lot of attention at the moment, it's not 1.0 either, and not to reach that point for years. Other languages like Odin, Vlang, and others that are flying under the radar might surprise us and may yet get more shine.

The only thing I care about is number of jobs available for language X.

Is there a list somewhere ranking languages by # jobs?

It would be interesting to see the evolution of those numbers and ranking over time, maybe on a month or quarter granularity.

I wonder where Perl, VB, assembly, SQL, R, Objective-C, and Matlab would end up on the list if they were included.

I would have guessed this and was glad to see my thoughts confirmed. Nice bit of ego support.

More seriously: it would be interesting* to do sentiment analysis on the articles and comments. A fair number of recent C++ articles on HN denigrate it, and any article about C++ inevitably generates a bunch of lazy anti-C++ comments. A post on "which languages do HN commenters like and dislike the most" would be quite eye-opening. I imagine rust would still be quite high, but perhaps the rankings would move a lot.

* interesting, yes, but not interesting enough for me to bestir myself to do this. I guess I'm just lazy...and not in the programming sense!

D is also missing.

poor scala, you deserve better

I love Tcl/Tk….

No surprise here, Rust users are physically incapable of going 10 seconds without mentioning that they use Rust.

Apparently not just Rust users but people who want to complain about it. Well done for keeping up Rusts numbers.

I've seen more complaining about Rust and its users on HN than I've actually seen Rust users.

People who get stuff done and are successfully using a technology rarely go around advocating for it.

Point being, Rust might be even more popular than what this post suggests. But then again that could go for any other language. Apart from Rust I also work with Elixir a lot and I know that most of its devs are not at all vocal; they're too busy being successful and don't need external validation -- they know the benefits of the tech and are using it to their full advantage.

I know that I do that, and I know others. Both with Elixir and Rust.

You mentioned Elixir (in a thread about Rust) and Rust a fair number of times to consider yourself "not at all vocal", so perhaps you are not a case in your point :)

The fact that I am mentioning the languages every now and then is not the same thing as constantly advocating for them.

Nuance, please. There is not only the 0% and the 100%.

I did not mean to offend, but simply highlight the nuanced contradiction in your comment for a good natured laugh on it. I haven't brought a single language up, and yet I'll admit to sometimes selling the benefits of one: it's only natural.

It's ok that you are vocal about them from time to time, but claiming how you are silent about that is simply untrue: you could have stopped after the first sentence, or not brought Elixir into the discussion at the very least.

I don't use Rust or make pro-Rust comments, but the pro-Rust comments don't bother me at all. They're respectful and substantive.

The anti-Rust comments, like yours, are irritating, off-topic, and not adding anything to the discussion. If you want to start a language flamewar, there are better places to do it.

I think the point is that many HN posts mention "I made X in Rust", whereas similar projects using Go or Javascript don't mention the language in the post title. So it does seem like Rust users tend to like vocalizing it.

I also think this study is slightly biased towards new and hip languages. For example, less people are going to vocalize that their new web app is made with Javascript, because that's already the assumption. But Elixir will almost certainly be mentioned

The pro rust comments on rust posts are great. The pro rust comments on every other post about any other language are not as great.

My favorite part is when it's a closed-source program so the language it's written in is barely relevant to its users

Any such Rust examples?

Discord: https://news.ycombinator.com/item?id=26227339

It's hard to say for most because, well, it's closed source. Unless they talk about it publicly like Discord, we won't know.

Discord is a commercial product that makes quite a lot of money though. Hardly a good example.

Had it been something like `ripgrep` -- which is open but yeah imagine it wasn't -- then I'd immediately agree with you.

And now I'm confused what point you're trying to make?

I mean, commercial products written in many languages are closed source and I don't feel that reflects negatively on any of them.

I'm not sure what point you're trying to make actually. You asked (as I understood it) for any Rust examples of closed source programs, for which I provided Discord. Now you're talking about how closed source does or does not "reflects negatively on any of them" which I never said anything about at all.

Ah, I was going to the root of your comment so assumed that you're posting it in support of the ridiculous claim that Rust users can't shut up about it.

Sorry if it's a wrong assumption then.

Couple off the top of my head: https://news.ycombinator.com/item?id=33150260

https://news.ycombinator.com/item?id=30921231

https://news.ycombinator.com/item?id=30952084 (this one’s funny because the team pulled the rug on their users after a month or two)

Thanks for following up.

When people point these out I feel obliged to mention they are hardly unique to any language.

I guess I asked because your comment was directly under a fairly ridiculous comment being fairly insulting to the Rust community, so I wondered if you want to point out deficiencies that are supposedly only present in Rust?

If not, my apologies for mentioning it.

I compile my Rust programs on Arch Linux.

I am disappointed you haven't moved to NixOS yet.

In all seriousness, I like reading about them all, because they do tend to move the needle in the right direction.

Reply on news.ycombinator.com