Following on from the great insights we got from last year’s typing survey, we’re going to be launching a new survey soon for 2025. I’m sharing the draft survey questions here for any feedback before officially releasing, please take a look and add comments. Also feel free to leave suggestions in this thread for how we can spread the word more broadly when it comes time to publish, a key point from last year’s insights was that it would be good to gather responses from a broader community of python developers, not just the typing community.
An important issue came up on that thread about last year’s survey, which I don’t see addressed here, namely that it’s unclear whether this survey is aimed at people who like static typing, people who use static typing, or just Python users in general.
If it is a goal of the survey to gauge interest in and appreciation of static typing overall (and not just relative interest in different tools, etc.) , I think that framing it as a “typed Python survey” will create a self-selection bias, as many people who aren’t interested in typing will ignore it. I’m not sure what the solution to that is, since just saying “Python survey” may be too vague. Perhaps some title or tagline that conveys neutrality towards typing would work (e.g., “Share your opinion on static typing” or “Do you like/dislike static typing”).
Relatedly, this comment from @bwoodsend last year seems unaddressed:
The survey includes a question for “why do you not use static typing”, but doesn’t include one for “why do you use static typing”. Such a question could have an answer choice like “to comply with requirements of my employer or projects to which I contribute”, which could capture the people who use static typing but don’t like it.
I’m just reading through the questions and thinking out loud[1]…
For the How many years of experience do you have using type hints in Python question, it’s awkward to answer if you’re an infrequent user. I could define it as either
For How often do you use type hints (static types) in your Python projects, I’m wondering how literally to interpret the “your” in cases where you do/don’t type-hint your own projects but do differently for collaborative projects. [2]
For If you don’t use static types, what are the reasons, could we replace the boolean factor/non factor choice with say a score out of 5? There’s some in there where I think Hmm, yes that does contribute and want to check it but I also don’t want to check it so as to avoid stealing weight from more significant factors. Similarly, with Where do you get help using Python typing.
For Which IDE do you mainly use, would it save you some post-processing of the results from the Other category to have a bucket response for IDEs that don’t have any Python and/or typed Python awareness (vim without extensions, nano, micro, pico, geany, notepad++, …, the world has no shortage of minimalist text editors)?
For How did you learn to use Python typing, maybe add trial and error and looking at existing typed code responses (I may be over-extrapolating my own experiences).
If you’re looking for questions, I’d personally be quite curious to hear how many people agree with the no non-typing related features should depend on type hints mantra that’s often used here to rebut a proposal where adding a type hint is the answer to some otherwise ambiguous behaviour.
I also want to say thanks for doing this pre-survey review. It’s always irks me when a question forces me to give a misleading answer. [3]
Yeah, I feel like a bit of a nit-picker for raising this one ↩︎
I’m currently still wriggling my eyebrows at the insights gained from one of the Python Developer’s Survey questions where I at least had to answer at random due to my real answer being N/A ↩︎
I’m kinda curious to see how often people explicitly use stubs-only packages like django-stubs, pandas-stubs, scipy-stubs, etc. So perhaps we could add a question like that?
this is a good point, we’ve tried to address this somewhat in the intro to the survey, and are drafting copy for when we share the survey to make it clearer that this is a pretty general survey, so whether you are familiar/keen on typing or not we are interested in hearing from you. You’re right that it’s hard to avoid a selection bias and calling it just “Python survey” is a bit too vague.
Thanks for the detailed points! In terms of how we format the questions we’re somewhat limited by what the survey platform will allow but we’ll look into what we can do! For some of the questions its also important we keep the same format as previous years so we can see year-on-year change. In terms of ambiguity of some of the questions (“how many years of experience..” etc.) we’ve tried to follow the example of larger surveys like the stack overflow survey which uses similar options. Ultimately all the questions are open to interpretation to a certain extent and it’s tricky to eliminate all possible misinterpretations
