The June 2019 iteration of the RedMonk Programming Language Rankings is brought to you by YLD.
YLD is behind many of the products and services you use every day. We create cutting edge technology and design capabilities that last beyond us. We’re experts in open source tech such as node.js, react.js, React Native, Kubernetes and design systems. Get in touch!
monkchips (James Governor): The purpose of this chat is to discuss the findings of the June 2019 RedMonk Programming Language Rankings, sponsored by our good friends at YLD. Were there any surprises, what’s going on, and what did we learn this term?
RedMonk publishes rankings every six months, and we felt it would be good to have a discussion this time around, and get behind the numbers and charts if we can.
We also asked for questions from third parties, and hopefully we can answer some of those.
rachel (Rachel Stephens):For those who didn’t see it, here is the aforementioned research and trends over time can be seen here.
kelly (Kelly Fitzpatrick): OK, to start off: any general impressions of this round of the Programming Language Rankings?
sogrady (Stephen O’Grady): My biggest takeaway from this round of the rankings was TypeScript. It’s the first new entrant into the Top 10 outside of a one quarter cameo from Swift since 2015. The growth behind the language has been impressive to watch.
monkchips: So about TypeScript – we got a little excited about the language’s progress last time, and it seems that interest was justified. What’s happening everyone?
sogrady: Two big factors in its success:
- JavaScript’s ubiquity and ascent to lingua franca status
- The (optional) safety offered by the language’s typing
rachel: I also think that the ability to refactor an app by mixing and matching TypeScript and JavaScript is a big part of the story.
sogrady: Safety is popular these days for many reasons, as Capital One (most recently) has emphasized.
Agreed, @rachel. TypeScript’s ability to intermingle with JavaScript allows it to ride the latter’s coattails, much like Kotlin with Java.
monkchips: Can’t overstate the truth of the point made by @rachel. The Javascript Typescript adjacency is very very powerful as a vector for adoption.
rachel: TypeScript’s focus on being tool-friendly is also huge. Tools are such an important part of developer productivity.
sogrady: Probably doesn’t hurt also that one of the most popular tools, VS Code, is written in TypeScript.
monkchips: Built by some of the same people that created the language…
It’s a really big win for Microsoft.
kelly: Agreed; and the compatibility of necessities like testing frameworks also seems key to me.
monkchips: Testing frameworks, great point. Any in particular in mind that have done a good job with TypeScript?
kelly: The lovely thing about Typescript is that usually–at least for unit tests–you can use whatever JavaScript framework you already love.
So it seems that TypeScript was big this time around; were there any surprises specific to the June 2019 rankings?
rachel: I feel like a turnover in the Top 10 is the biggest news. The top programming languages are relatively stable given that our metrics are accretive over time.
monkchips: Agreed. I did get this tweet which might be worth discussing
HCL’s position on the chart is surprising, I’m guessing that’s because users on StackOverflow are more likely to tag “terraform” than “HCL” in their posts?
— Jon Topper (@jtopper) July 30, 2019
To be honest I had never really thought of Hashicorp Configuration Language’s place in the rankings at all before.
But Hashicorp has talked about it a bit recently, and it was notable that GitHub chose to use it with Actions.
It seems like an interesting outlier – quite a bit of GitHub activity but very little in Stack Overflow
sogrady: Looks like it clocks in at #62 this time around. Will be interesting to see if it rises.
rachel: I imagine the hypothesis around adjacent tags has merit on the Stack Overflow side, though I’d have to go look to confirm.
monkchips: GitHub Actions feels like such an interesting potential context and vector for further adoption. Actions workflows are written in HCL
sogrady: Ah, that’s interesting: HCL was #70 in January, so #62 is a fair jump, though admittedly the distinctions are less meaningful that far down.
rachel: We often get pushback about the rankings for communities that gather and collaborate outside of Stack Overflow.
kelly: We definitely had some questions around the processes used to create the rankings.
rachel: This is a similar objection. Our process is to correlate trends from two publicly available data sets (namely GitHub and Stack Overflow) to try to see what it tells us about usage and discussion.
There are many communities that collaborate in places other than Stack Overflow, and we simply can’t expand to all possible data sources (partially because not all data sources have publicly facing data, but also because we the process would not scale.)
The same logic applies to tags. We try to catch the primary tags associated with a language, but we can’t expand to all the adjacent tags.
kelly: We did get one question around YAML (that begs the larger question of what is considered a “programming language”):
Should YAML be considered a programming languages? And if it was, where on the rankings would it fall?
Also, when is this chat and is it open to an audience? Would love to join.
— Gareth Rushgrove (@garethr) July 30, 2019
monkchips: YAML is currently a language folks love to hate, but it’s utterly ubiquitous. It’s definitely not a “computer science” language, but it’s a get shit done language. Folks like Pulumi and Atomist are actively arguing its use is harmful.
rachel: The reason it’s not included in our rankings is because GitHub is not including it in their data.
sogrady: Right. We try to make as few editorial decisions as possible, so until GitHub’s Linguist project considers YAML a language it’s not eligible for inclusion.
monkchips: We might have to talk to GitHub about that!
So…I have a question about Python. It’s there, well-established in the top three now, it’s become a lingua franca for data science, it’s proved itself general purpose. Or has it? I have been noticing some chatter from folks recently that are seemingly arguing we might be at Peak Python, because while it has found a place in AI/ML, it’s not really nailing other opportunities. That is, might it face stronger headwinds going forward?
sogrady: I think they’d say: PR’s welcome!
And we'd welcome pull requests for new languages if they meet the in-the-wild usage requirement: https://t.co/2N6jwP0Ts5.
— Paul Chaignon @[email protected] (@pchaigno) July 21, 2019
Re: Python, there’s so much of it out there in so many places for so many things that it’s not going anywhere for a while. That said, there could still be something to the Peak Python argument because it lacks the clarity of purpose and design that, say, an R has, but isn’t likely to challenge JavaScript/TypeScript for ubiquity.
monkchips: Right.
sogrady: Which raises the question of what its path is moving forward, and what opportunities it’s positioned for. Still, I wouldn’t bet against Python. It’s good at so many things.
monkchips: Guido van Rossum stepped aside as language overlord, and without a benevolent dictator, arguments about language direction are likely to increase.
rachel: But it’s the Python community, so the arguments will still be welcoming.
kelly: Also re: Python: in an informal conversation with some .NET and Java devs, there was bafflement over how high Python fell in our rankings. At the same time, almost everyone I know that is starting to code starts with Python.
monkchips: @sogrady: Agree wholeheartedly plenty of life left in the old dog yet, but I have definitely been noticing an uptick of negative chatter lately from long time users of the language
sogrady: @rachel as an aside, Python is welcoming, but has nothing on Haskell
“what happens when you troll the haskell community”: https://t.co/emW92JjiNj
— steve o'grady (@sogrady) May 9, 2014
monkchips: So what’s Haskell doing this time around?
rachel: It’s been hanging steady at 19 since 2017.
kelly: Switching gears, is there anything that you see potentially changing in the process that we use for the language rankings? Any fine-tuning that we should consider (or changes that might be driven from external factors)?
rachel: Selfishly, I would love if the answer was no.
sogrady: Same. Process changes are the worst, and I’m not even the one who has to bear the brunt of it, @rachel is.
rachel: Realistically, the world always changes. Data format/availability changes are occasionally inevitable. And then there’s the questions that come around whether we need to adapt the process overall to more communities, etc.
Right now we’ve found GitHub and Stack Overflow to have large developer traction and publicly available data. But as we pointed out above, these sources are far from giving us a universal picture of language use.
monkchips: Evolution is good, it really might be interesting to look at GitLab and Bitbucket data, given their enterprise footprints, but obviously that’s making work for you, and would potentially be a “breaking change” on our timeline.
rachel: People ask often about expanding our source data. I feel like that might add interesting color, but it also would make views of languages over time less relevant when you do that big of a process change. It’s a balancing act.
kelly: Are there any changes we are considering in how we present the information as we currently gather it? @monkchips: you have had requests for information on languages that do not end up in the publicly available Top 20:
@monkchips are the language rankings past 20 public, or part of paid access? Didn't see a full list from perusing. Just wondering where coldfusion and Delphi are…
— Ken (@kenhorn) July 29, 2019
monkchips: For Ken Horn the short answer is no, we don’t have secret list below the fold for paying customers. Is he planning to write some ColdFusion?
sogrady: @monkchips there is no secret list, but we have periodically looked at downranked languages for clients who want to know how something is trending outside of the Top 20.
I can’t see the rankings as we currently run them changing for that reason; we have almost a decade’s worth of continuity to preserve. But it’s certainly possible that we could see additional ranking types added at some point in the future.
monkchips: We could have a language rankings FORK!
kelly:
sogrady: Haha yeah that’s not happening.
rachel: That’s the beauty of open data sets. You are welcome to fork, @monkchips.
sogrady:
I have to admit that @rachel: having the most epic troll ever is something I did not see coming.
kelly: We are close to time, so one final question:
Any predictions on what we might see in our 2020 rankings (either in the rankings themselves or in a fork from James)?
rachel: I predict James will not fork.
sogrady: I’m curious to see what the future holds for Go. On the one hand, it’s very well regarded technically and a foundation for numerous large, critical projects. On the other, it’s use cases are relatively narrow and there has been a fair amount of community dissent over input, or rather the lack thereof, in its evolution.
It’s never placed higher than #14, IIRC, and I don’t expect it to see that again anytime soon.
rachel: I’m hoping Rust cracks the top 20 at some point.
monkchips: HCL and we should really encourage @GitHub to track YAML.
sogrady: Yes, we’ll have to keep an eye on HCL. Rust I expect to hit the Top 20 if not next run, the one after that. The trendline is pretty clear, and if this comes true Rust would get an immediate boost:
Microsoft is exploring using the Rust programming language as an alternative to C, C++ in the name of better security: https://t.co/nJZmcOPWXe (by ZDNet's @campuscodi)
— Mary Jo Foley (@maryjofoley) July 17, 2019
kelly: And I vote that we do another language rankings chat in 2020–this has been great!
rachel: Indeed! Thanks, team!
No Comments