I gave a lecture on LSST data products at the DE School in Oxford, and ran an interactive poll during the talk (using PollEverywhere). The full results are here:
Of most interest to us were answers on how people get their Python (>50% get it with Anaconda), and which version they’re most comfortable with (still 2.7, by a surprising margin). The nice thing about this is that it’s a relatively unbiased sample (of students/postdocs interested in LSST’s Dark Energy science :)). It’d be cool to run a similar thing on an even broader sample (e.g., during one of the AHM plenaries). The audience loves it, and we get feedback fast (I’m asking about getting us access to pollev for larger audiences).
Also of interest: Empire Strikes Back is still the most popular Star Wars :). Unfortunately, I didn’t get a chance to ask about the best Doctor (Tennant, obviously)…
Or best episode… Blink obviously
On a more serious note, might be good to distinguish between OSX and Linux for preferred Python distribution method. I would be surprised if people are using system default Python on OSX.
Are you sure this is a good statistical sample when there are so many people named “andrew” and “david”? Isn’t it possible that the numbers are skewed by a couple of people polling multiple times?
Tom Baker for favourite Doctor. That scarf.
and so many fans of the prequels…
Going back to the serious question. The thing about python2 is one of motivation. People don’t feel any need to upgrade because their code works and some of the things they depend on don’t work on python3. Why change?
As developers of library code and people writing code that still has to work in 2030, we have to worry about Python3. If that motivates people to switch away from legacy python sooner rather than later then that’s a bonus.
See also: https://github.com/astropy/astropy-APEs/pull/13 proposing the time line for Astropy dropping legacy python.
I wondered if anyone would ask :). Yes, that’s a know issue (for that particular question) – I’m looking at you, @connolly… The others were limited to one answer per session.
yes… the Chicago voting system might have been employed for that question