[Python-Dev] Python 2.x and 3.x use survey, 2014 edition

[Mark Roberts]

On Mon, Dec 15, 2014 at 11:30 AM, Chris Barker <chris.barker at noaa.gov>
wrote:

> Are you primarily writing packages for others to use? if so, then yes. But
> I wonder how many people are in that camp? Don't most of us spend most of
> our time writing our own purpose-built code?
>
> That might be a nice thing to see in a survey, actually.
>

So, I'm the guy that used the "hate" word in relation to writing 2/3
compliant code. And really, as a library maintainer/writer I do hate
writing 2/3 compatible code. Having 4 future imports in every file and
being forced to use a compatibility shim to do something as simple as
iterating across a dictionary is somewhere between sad and infuriating -
and that's just the beginning of the madness. From there we get into
identifier related problems with their own compatibility shims - like
str(), unicode(), bytes(), int(), and long(). Writing 2/3 compatible Python
feels more like torture than fun. Even the python-future.org FAQ mentions
how un-fun writing 2/3 compatible Python is.

The whole situation is made worse because I *KNOW* that Python 3 is a
better language than Python 2, but that it doesn't *MATTER* because Python
2 is what people are - and will be - using for the foreseeable future. It's
impractical to drop library support for Python 2 when all of your users use
Python 2, and bringing the topic up yields a response that amounts to:
"WELL, Python 3 is the future! It has been out for SEVEN YEARS! You know
Python 2 won't be updated ever again! Almost every library has been updated
to Python 3 and you're just behind the times! And, you'll have to switch
EVENTUALLY anyway! If you'd just stop writing Python 2 libraries and focus
on pure Python 3 then you wouldn't have to write legacy code! PEOPLE LIKE
YOU are why the split is going to be there until at least 2020!". And then
my head explodes from the hostility of the "core Python community". Perhaps
no individual response is quite so blunt, but the community (taken as a
whole) feels outright toxic on this topic to me.

Consider some statistics from Pypi:
- 13359 Python 2.7 packages
- 7140 Python 3.x packages
- 2732 Python 3.4 packages
- 4024 Python 2.7/3.x compatible packages
- 2281 2.7/3.4 compatible modules
- 9335 Python 2.7 without ANY Python 3 support
- 11078 Python 2.7 without Python 3.4 support
- 451 modules 3.4 only packages
- 3116 Python 3.x only packages
- 1004 Python 3.x modules without (tagged) Python 3.4 support

Looking at the numbers, I just cannot fathom how we as a community can
react this way. The top 50 projects (!!) still prevent a lot of people from
switching to Python 3, but the long tail of library likely an even bigger
blocker. I also don't understand how we can claim people should start ALL
new projects in Python 3 - and be indignant when they don't!. We haven't
successfully converted the top 50 projects after SEVEN YEARS, and the long
tail without 3.x support is still getting longer. Claims that we have
something approaching library parity seem... hilarious, at best?

I suppose what I'm saying is that there's lots of claims that the
conversion to Python 3 is inevitable, but I'm not convinced about that. I'd
posit that another outcome is the slow death of Python as a language. I
would suggest adding some "community health" metrics around the Python 2/3
split, as well as a question about whether someone considers themselves
primarily a library author, application developer, or both. I'd also ask
how many people have started a new application in another language instead
of Python 3 because of the split.

-Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-dev/attachments/20141215/3431694e/attachment.html>