Creating a perl developer survey
Having seen the Javascript developer survey, I thought that Perl should have one (there was one in 2010).
Here's my current list of questions. I've got these in a Google docs form, but won't show the options for each question.
- Is Perl your primary development language?
- What level of Perl experience would you give yourself?
- How long have you been programming in Perl?
- What version(s) of Perl do your systems run on?
- Is programming in Perl part of your daily job?
- How do you find reusable code, libraries and tools?
- Have you released any code to CPAN?
- Have you written any code in Perl 6?
- What is your preferred OO framework?
- What development environment(s) do you use for Perl?
- What framework, if any, do you use for web development?
- Other than Perl, what are your primary development languages?
What other questions should be on the list?
If you program in Javascript, go fill in that survey as well.
Perl hacker since 1992.
I think this is a great idea and this is a good start. I'd personallt be interested in how people view the future of Perl, what current and upcoming Perl projects are exciting or important and what things they wish the community would do better. Also, some thoughts on what they might wish to have had available when they started with Perl would be nice to have for those of us working on Perl documentation and tutorials.
Re "what version(s) of Perl do your systems run on?"... you should clarify whether you mean live systems or development. I've got 5.6 and 5.8 installations hanging around for development, but the oldest version of Perl I use for actual live deployments in 5.10.1.
Don't reinvent the wheel please. I was planning on running a perl survey in 2013. There have been a few surveys around recently and I wanted there to be a bit of a lull.
If you want to help, grab https://github.com/singingfish/Data-PerlSurvey-2010 - and look at the questions list and revise/push/pull request as you feel appropriate.
Next up, the design, data curation and analysis for surveys is not trivial, and the R code for that survey does some fairly involved stuff. Sticking to a reasonably similar set of questions can make the data mangling and analysis code reasonable to reuse. It also provides the facility to make comparisons with the previous run. One loses a lot of information by not building on prior work.
I have hosting for the next survey organised (via a sponsor), and it makes a big difference to have to not worry about that side of the infrastructure.
But I would love for contributions. Just please don't reinvent a wheel that took quite a lot of effort to get somewhat right the first time. Grab me on irc.perl.org (kd) or twitter (kd__) to discuss in real time.
Also looking at the reports on the 2010 survey and evaluating what questions worked well, which were incomplete and which could be deleted would also be very useful in refining the instrument. Bear in mind I got the grant to make a repeatable instrument, so the time investment in running the survey in a similar manner again using the same infrastructure should be about 10fold less than the original effort required.
You can expect around 4-6000 responses from running a developer survey properly with this group of people, and generally the data quality seems quite high. So yeah it may look more complex initially to go with the existing survey instrument, but in terms of understanding the findings and getting something of high quality out in a reasonable time period, it's a much better idea.
Excellent! -Glad to hear from you, and to hear that you're planning to run this again. I'll get in touch offline. I agree with all the points you made.
I would like to know what platforms Perl programmers are using. I write most of my Perl scripts on and for Windows these days and I'm curious how many other Windows Perl programmers are out there.
Here you go. Back in 2010 the answer was operating systems ... and best windows distributions. So for the 2013 survey, we should add dwim perl. Any others?
Thanks for the links. I'm pleasantly surprised by how high Windows is in the percentages. That tells me that CPAN smoke testing on Windows is more important than ever.
Please add Citrus Perl to the survey.