The 2021 Julia language survey has completed

Julia users and developers are invited to complete the annual Julia User & Developer Survey. Click here to respond — make your voice heard and be counted as part of the Julia community. The survey is available in English, Chinese (also mirrored on qq), Japanese, and Spanish.

As usual, findings will be presented at JuliaCon. You can check out the results from last year’s survey to see what this looks like.


Some quick feedback: I have started using Julia in 2021 and it is regarded as the upper limit in the question, but the input field does only allow a maximum value of 2020.


Ah, shucks, I had thought we fixed that. Thanks for the report.


Done! :slight_smile:

Also a bit funny is that 2009 is the stated lower limit but allowed are actually only values from 2012 onward… I thought I better report this so Jeff, Stefan and Viral don’t struggle with the survey.


Also, for the first two questions, would it be possible to preselect “none” and “don’t know / no opinion”? It takes quite a bit of dedication to tick the same box for most languages, and you risk putting people off by forcing them to do so.


Default answers on surveys are always tricky and need to be handled with extraordinary care. I was able to navigate the table fairly effectively from the keyboard with tab and arrows, but I agree it’s a slog of a first question!

Now that the survey is live we won’t be making any changes (excepting the 2021 bug which is now fixed). Some things need to stay consistent to allow for year-over-year comparisons, but you can keep your eyes out for the request for feedback for next year’s survey. Here was this year’s thread: Feedback on Julia User & Developer Survey 2021


Under academics, for courses, I feel there should be an option for “Other” and one can fill in their course of study. I study Systems Engineering and I had to pick “None” because it wasn’t on this list.

1 Like

Was surprised to see one-based indexing as one of the possible answers in both of the technical aspects questions, while the (arguably more important) column-major order is not there.


Totally agree. My approach was to simply only answer on languages that I had some input on and ignore the ones I didn’t. So “default” answers will occur if you like them or not.

Seems to me that this would have been solved by completely excluding the “None” or “Don’t know, no opinion” columns in those two first questions and adding an explanation along the lines of “ignore languages you don’t use or lack an opinion on”.


It is too bad that this years form is equally confusing and (IMO) strange (The 2020 Julia language survey is live now - #19 by fredrikekre). Maybe next year :slight_smile:

Perhaps it would also be worth clarifying who is performing the survey. Is it JuliaLang or JuliaComputing? I couldn’t find this information in the OP or in the survey itself.


as a nice side effect, I learnt the existence of a few new programming languages :laughing:

The list in the question “what are your favorite packages” seems a little arbitrary to me, I was surprised that e.g. Literate.jl is not there. Maybe the question should have an option “others” and give the user the possibility to list their favorite packages? I know open questions are always a pain in surveys, because they mean a lot of manual processing, but maybe the manual work could be minimised with clear instructions such as “list your favorite packages separated by a comma”, or an add bottom that adds a text field for each package the user wants to write. This and the assumption people can spell their favorite packages should greatly reduce the manual work :smiley:


I agree. Ideally, there would be a drop down menu with auto completion that’s linked to julia hub for a complete list of correctly spelled package names. Ideally…


I’m missing nano from the editor list! :smiley: There are dozens of us! Dozens!! (even if it’s only for really small and quick codes - otherwise I use VSCode).


Yes, this is run by Julia Computing.

Oh it’s very arbitrary. The best way to “vote” for your favorite packages is to star them on GitHub. I did notice that we didn’t include all of the 10 most starred packages (ranked by stars) after deploying this, though, and in the future we’ll be sure to at least include those.


Glad to hear this! As a Gen contributor I was surprised to see that it wasn’t listed this year.

1 Like

Once again (as per every past year, and as per every year i am griping) we have written the nonsensical sentence:

features you like MOST about Julia? Please select all that apply.

MOST and select all are mutually exclusive.
Either it is select the one feature you like most,
or it is select all features you like (or all you liked significantly or some such)

I guess because we screwed up that the first time, and we want to keep consistency for cross year comparison.


The options for question 11 feel like they should be worded differently. The “Created or developed” vs “Contributed to, but did not create or develop” make it sound like you can only create a package or contribute to an existing package, but not do both. Specifically, I don’t understand what is the difference between “developed” and “contributed to” in those options.

Personally, I would change it to a “select all applicable” question, with “created a package”, “contributed to existing package(s)” and “none of these” as options.


This post was flagged by the community and is temporarily hidden.

Hi, Big Data is completely missing in the survey. Julia is extremely nice when you have to deal with high amount of data, using JuliaDB. It is way easier than using the Hadoop big guns in many cases. Also, it avoids the performance issues of Python based solutions when you start doing some custom operations.