Is there any evidence for this supposed trend? Or will these “many” programmers remain merely a vague hand-wave?
My own Python code, over time, became more and more functional in style until eventually I jumped ship to Clojure and Yeti (and over the past week, F#).
[1]: http://github.com/mbrubeck/outline-grep
But I'm certainly ready and interested in moving toward a more functional language -- I've always been interested in the Lisps, and have always tended to write code in the functional style, where possible (even though my first language was Java...)
Then again he's programming in a hot bed of programmers in that well known programming mecca of Hamilton, New Zealand. Oh, wait...
i think you must mean "false", then.
But there wasn't anything in Python that I found compelling so I never came back.
As a scripting language, Ruby seems so much more compelling, despite it's friendliness seeming to doom it to eternal slow bugginess. I'd go back to Ruby before Python.
But I'm more likely to add Python support to an app I'm developing, since I know it's more sensible.
Python's functional aspects seem chronically undeveloped so leaving it for a real functional language makes sense. But functional approaches seem limited in the "real world".
Then again, Python doesn't have rake.
Point the first: Python 2.x to Python 3.x if you already have to make a transition, there is more impetus to make a transition to something that suits how you actually work.
Point the second: itertools is partly a catalyst, partly an indicator. Once you get fluent with it, it illuminates pain points.
Point the third: There is a new batch of functional languages that support the style that is a pain point, as well as fixing other issues that hurt in python.
Lastly, not all python programmers will want to make the shift. But the ones fluent with itertools are more likely to be the subset that will.
I did in fact switch during the 2.x->3.x transition, though I haven't given up on Python entirely - I still know the language, I'm fine with working in it, It's just not my first choice anymore. If there were one factor (and there isn't, my switch to Lua was gradual* ), it would be GvR's insistence that Python will never get tail-call optimization, because it's "unpythonic". (http://neopythonic.blogspot.com/2009/04/tail-recursion-elimi...) That has major implications for the expressiveness of the language.
* I think what happened was that I had been reading Lua's source to see how a small, clean implementation of a bytecode compiler worked, and along the way discovered that it suited my preferred development style better. Lua feels like a minimalistic hybrid of Python and Scheme, with the "good parts" of Javascript (prototypes, object literal notation) harmoniously integrated, but designed with embedded and multi-language development in mind. Also, the whole language can be installed with a DLL smaller than SQLite (!), which makes deployment rather easy.
When I say in this position, I mean that I would like to switch but won't in the next future (too costly).