I think there's an underlying misunderstanding in all this discussion. I don't think anyone ever thought GH or SO were perfect mirrors of the language usage of the industry. That would be gullible. But, you know, they're the best we have.
Knowing that nothing will even be a perfect statistic representation, should we flat out stop saying anything about it?
On GitHub, JavaScript is the most used language by a long shot.
On Stack Overflow, it's the most discussed, by a long shot.
Among package registries, npm is the largest. By a long shot.
I won't say that three clues make a proof, of course. But we're not talking about "JavaScript is 15% ahead of LanguageX", but rather an order of magnitude. I think it's pretty reasonable to draw conclusions, although without exact numbers.
Stack Overflow's co-founder Jeff Atwood, who is not a JavaScript developer (rather a .NET dev) stated his famous law in 2007: "Any application that can be written in JavaScript, will eventually be written in JavaScript." And in 2007 we didn't even have a lot of things that made JavaScript so popular: Chrome, Node.js, Angular, React, ...
Personally, I find Stack Overflow's usage quite interesting because - assuming no language/environment is significantly more difficult than another - it's about the struggles of everyday's users, and their direct experiences.
On the other hand, I feel you're playing as the devil's advocate here. Don't mistake me, it's often an important role, as it can lead to find trivial oversights. And yet, even with all your considerations, I fail to see how those aspects can mean an effective bias.
For example, the spoken language. Granted that most of the surveys and Q&A sites are in English, so developers who don't speak English (well) are less likely to use them, it still doesn't explain why this should affect the numbers. Does the English-speaking community lean towards JavaScript? If yes, how? And how much?
What about SVN and TS? Is there a reason to believe that private repositories contain, proportionally, less JavaScript than GitHub (at least the part that's public)?
Your reference to Hackerrank is indeed interesting, but: a) 2016 is three years ago, which could be considered a long time in IT, and b) interviews are influenced by the market's offer.
The application field, on the contrary, is definitely a factor. Of course the gaming industry will push for low-level, high performing languages such as C; and the scientific community is shifting towards Python thanks to its well-crafted libraries. I don't think anyone will even contest these points, as there are clear reasons.
But we're looking at the overall picture here.
Sorry for the lengthy post - but I want to let you know that, in the end, you actually made think and ponder 🖖
Given your response, yes I would never use either of those three statistics or the combination to make a blanket statement that X is the most popular because, as you state, we can't. I think it's misleading and uninformed.
We can say plenty using those three stats without making assumptions and call them as if they're facts :).
I can go into the specific questions you ask later, when I have the capability to give you an articulated response, but for now I'd would say: I'm glad I made you think about it, yes I am partially playing the devil's advocate because I wánt people to think about it and yes: I think there is far less JS comparatively on non-github svns, which ties in with less open source, less packaging and more enterprise and closed source software.
Do I think JS will keep growing? DEFINITELY! Will I use SO and GH as the source to determine that it's the most popular language? Personally I will never for the reasons I've mentioned so far, until I can be more certain that they are a good reflection of the programming world as a whole.
I think there's an underlying misunderstanding in all this discussion. I don't think anyone ever thought GH or SO were perfect mirrors of the language usage of the industry. That would be gullible. But, you know, they're the best we have.
Knowing that nothing will even be a perfect statistic representation, should we flat out stop saying anything about it?
On GitHub, JavaScript is the most used language by a long shot.
On Stack Overflow, it's the most discussed, by a long shot.
Among package registries, npm is the largest. By a long shot.
I won't say that three clues make a proof, of course. But we're not talking about "JavaScript is 15% ahead of LanguageX", but rather an order of magnitude. I think it's pretty reasonable to draw conclusions, although without exact numbers.
Stack Overflow's co-founder Jeff Atwood, who is not a JavaScript developer (rather a .NET dev) stated his famous law in 2007: "Any application that can be written in JavaScript, will eventually be written in JavaScript." And in 2007 we didn't even have a lot of things that made JavaScript so popular: Chrome, Node.js, Angular, React, ...
Personally, I find Stack Overflow's usage quite interesting because - assuming no language/environment is significantly more difficult than another - it's about the struggles of everyday's users, and their direct experiences.
On the other hand, I feel you're playing as the devil's advocate here. Don't mistake me, it's often an important role, as it can lead to find trivial oversights. And yet, even with all your considerations, I fail to see how those aspects can mean an effective bias.
For example, the spoken language. Granted that most of the surveys and Q&A sites are in English, so developers who don't speak English (well) are less likely to use them, it still doesn't explain why this should affect the numbers. Does the English-speaking community lean towards JavaScript? If yes, how? And how much?
What about SVN and TS? Is there a reason to believe that private repositories contain, proportionally, less JavaScript than GitHub (at least the part that's public)?
Your reference to Hackerrank is indeed interesting, but: a) 2016 is three years ago, which could be considered a long time in IT, and b) interviews are influenced by the market's offer.
The application field, on the contrary, is definitely a factor. Of course the gaming industry will push for low-level, high performing languages such as C; and the scientific community is shifting towards Python thanks to its well-crafted libraries. I don't think anyone will even contest these points, as there are clear reasons.
But we're looking at the overall picture here.
Sorry for the lengthy post - but I want to let you know that, in the end, you actually made think and ponder 🖖
Given your response, yes I would never use either of those three statistics or the combination to make a blanket statement that X is the most popular because, as you state, we can't. I think it's misleading and uninformed.
We can say plenty using those three stats without making assumptions and call them as if they're facts :).
I can go into the specific questions you ask later, when I have the capability to give you an articulated response, but for now I'd would say: I'm glad I made you think about it, yes I am partially playing the devil's advocate because I wánt people to think about it and yes: I think there is far less JS comparatively on non-github svns, which ties in with less open source, less packaging and more enterprise and closed source software.
Do I think JS will keep growing? DEFINITELY! Will I use SO and GH as the source to determine that it's the most popular language? Personally I will never for the reasons I've mentioned so far, until I can be more certain that they are a good reflection of the programming world as a whole.
Also: thank you so much for taking the time and writing it like you did. ❤❤❤