When will python3 be abandoned?

When will python3 be abandoned?

Other urls found in this thread:

perl6intro.com/
jstorimer.com/blogs/workingwithcode/8085491-nobody-understands-the-gil
twitter.com/AnonBabble

it's on planned obsolescence, so when python4 comes out.

When will python2 be abandoned?

never, it'll outlast the python3 meme

python2 is basically latin, its lack of updates is precisely why its used so much

Give it a few more years for Python2 to die. Python 4 will have backwards compatibility with Python3

>Give it a few more years for Python2 to die.
it wont ever die. its the only option on the shitloads of centos 6 and even 5 server parks out there, and python2 just werks. there is no need. hell, in my field the main tool still has no plans to support python3.

im almost certain redhat will gather a consortium by 2020 to continue support for python2. its literally latin, and thats important for webdev hipsters to understand.

This bullshit reliance on older languages and older software is why technology will never improve.

Python2 does literally everything you need. New languages are not what improves tech, it's new algorithms and faster hardware.

This bullshit obsession with fad languages and fad software is why technology will never improve.

You havnt been around long enough to understand what stability is. Grownups understand that fashion is something for sub-intelligent people.

By the way, the language you speak here is about a millenium old? Isnt it time you upgrade?

Porting form Python 2 to Python 3 is really not hard.
My company doesn't use Python 2 at all anymore.

C is half a century old now. Its not going anytime soon.

For trivial webdev its not hard no.

>By the way, the language you speak here is about a millenium old? Isnt it time you upgrade?
inb4 Sup Forums accusations

The new async stuff is what will probably get the webdev hipsters to move on. It's been all the rage in python frameworks to use "greenlets" and other bullshit terms.

The problem lies with Python and it's GIL though. If they'd actually managed to get rid of the GIL issues in python people would've moved on to Python 3 now. Instead you get a lot of handwaving about better unicode support, which then turns out to be the same shit everyone had with unicode, but now nicely separated from binary.

I guarantee you that python 2 will still be available on distros in 2020 though. Simply because of the fact that most people can't be assed to port their code, including most distro maintainers.

When will python be abandoned?

When Perl 6 replaces it

That or when Guido Van Rossum is caught fucking a parrot up the ass in his Google office

If you actually followed Python development, you'd know that there is an active effort to remove the last remaining bits of the GIL and that work is currently approaching production quality.

Python3.5 is indeed starting to offer significant benefits over 2, and personally I did make the switch earlier this year.

However, it's very little, and very late. The main number one problem Python has is speed, and that doesnt seem to be even planned to be solved anytime soon.

Its like perl5 and C before that: they're here to stay. Perl6 wont change that, Perl 6 wont, and Python3+ wont. Python2 made it to the hall of fame, its Guidos fault for not recognizing that breakage over nearly nothing was a retarded move.

Good for them, then finally the "async is better than threading" wankery can stop and people can drop the whole "multiprocess is the way to go" meme while they're at it too

It's created a node.js tier of mental retardation in webdev circles, and I'll be glad when it's gone.

But user async IS better than threading for most of the applications web devs will create

You can actually get decent performance out of Python if you start messing about with C. Numpy is a fairly good example, and it's ridiculously easy to get python to talk to a C library

But yeah, I agree that it's too litte too late.

And using Numpy or other external libs basically mean you dont care about the Python itself anyway (Numpy API is identical in Py2/3). So even more reason to take the Python version thats available on every server on the planet instead.

Yes but you can write async code in non retarded languages and suddenly the hilarious BS that Node forces down your asshole because "mun async" seems more retarded.

My PokemonGO bot runs on Python2 ebin. It can't be killed.

It's not possible to remove GIL from CPython without hindering single threaded performance, breaking C API, and significantly complicating an interpreter. All of this is something BDFL isn't going to approve.

I guess my point is that async and threading are like peanut butter and jelly. They work individually, but they're two different things that work well together.

Right now the mindset seems to be that it's either the one or the other. Over the years I've been through this kind of reasoning before in C, and then in Java, and now it's here again in Python.

Give it another 5 or so years and we'll be having this discussion again with another language, and we'll be rolling our eyes again.

But guys better print function, srsly.

from __future__ import print_function

>not hard
>all those libraries only for version 2

Most libraries have been ported to Python 3.
Regarding the few that weren't, well get your hands dirty and port them yourselves.
Someone has got to do it.

good, most of your new car has passed safety inspection, most of your food has not spoiled, most of the tumor was removed. you do the rest!

Why is python2 considered latin in comparison? There's like 4 tiny differences

Why is GIL such a big deal? Most of the time threading is used to avoid blocking on networking.

>Why is python2 considered latin in comparison? There's like 4 tiny differences
It's not the differences, it's the rate of change.
Python 2 is set in stone, if you know how to code perfectly in it now you'll know how to code perfectly in it forever.
Things are changing in python 3, new things are being introduced. The best or 'most correct' way of coding something might change every month, core principles might change too. This can really fuck with your programs, especially if you're trying to distribute it to other computers/servers that may or may not have the latest python 3 update.

i still believe in perl 6 ;_;

You could say that the existence of Python3 made Python2 into Latin.

if you are a peasant webdev

It isn't dead or anything. It's just not done yet.
perl6intro.com/

It's pretty sweet looking. It has variable binding and a type system with support for runtime checks similar to Eiffel's Design by Contract or Racket's contract library. The first feature makes MVC possible without the observer pattern and the second is arguably more useful than the static type checkers from the Miranda/Haskell/ML family.

python2 code written a decade ago still works, and will work in the decade to come.

You're missing the point. GIL was good enough for CCP to make a successful MMO out of Python. The GIL is good enough for Rackspace.

GIL is usually not such a big deal, because typically like you say, when you're dealing with threading you're often waiting on I/O anyway. My personal grievance with it is that Python is becoming ever more popular in academics and R&D where often instead of I/O bound applications you're dealing with CPU bound algorithms.

Now, you can argue: right tool for the right job, do it in C/C++/Java/other_preferred_language, but most of these researchers aren't computer programmers. Far from it actually. Their closest experience is usually dabbling about in matlab, but the cost of matlab is often prohibitive which is why python and numpy, scipy etc are becoming so popular and you get those python distributions like anaconda which try to include everything but the kitchensink. Not to mention that things like Jupyter make it terribly convenient for people without a technical background to start doing stuff in python and get results quickly.

This is where the problem with the GIL often comes in. The moment your dataset starts becoming bigger than the mousefart most researchers test their code on you need to start looking at a way to do multiple things at the same time.

You are missing the point: not everybody writes user interfaces with Python

>Python 2
>Python 3
Why that's even a thing to begin with?
couldn't they just keep updating the same language instead of making this compatibility nightmare since they can't aim for speed anyways?

Fucking retards, this is why C++ will always be superior.

And that's when the fun really begins. You get to pick between threading and multiprocessing, but the way to share data between processes in multiprocessing is an obsolete shitfest. If you're lucky you can rewrite the algorithm in such a way that there's very little need for sharing memory resources and then it's not a big deal if you use multiprocessing, but in many algorithms that's not the case.

Now, another option is just to rewrite the algorithm in C/C++/other_language and keep the python code as a reference prototype implementation, but then you need to have a good justification for it since obviously costs time and money to do so.

If I remember correctly, CCP made Eve Online in Stackless Python back in the day, which is a slightly different beast when it comes to threading and it uses the concept of microthreads and coroutines (think like LUA). The GIL is less of an issue there because they simply don't use threading.

I don't know if that's still the case today with Eve, but I clearly remember the hype back then about Stackless Python going to solve all of our woes... (it didn't)

Design by comitee is a real real bad thing..

>just_wait_OP_will_surely_deliver.tiff


This. Once you remove the GIL you're in for a lot of trouble.

This article is about the Ruby GIL, but it's the same problem for python:

>jstorimer.com/blogs/workingwithcode/8085491-nobody-understands-the-gil

Why do you want it?

Name some.

pyROOT

>dis niggah dont even particle physics

Haha I hate myself

that's because C is capable