Python 2.5 adoption

[Joseph Turian]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

So I guess the question is, does anyone have a sense of what percent
of python users don't have 2.5?

Thanks,
Joseph

 

[Mike Driscoll]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

So I guess the question is, does anyone have a sense of what percent
of python users don't have 2.5?

Thanks,
Joseph

I think it depends more on what you want to do. If you're distributing
the software, you can just "freeze" it and make binaries and then it
doesn't matter. Or if you use Python at your business, you can do what
we do at my workplace: Put Python on the network and run all the
scripts from there.

Currently, we have 2.4 on our network, but I think we can upgrade it
to 2.5 without breaking anything. I develop in 2.5 and just put the
finished products on our network and they usually "just work". But I
have yet to find good use cases for some of the cool whizz-bang extras
of 2.5, so I haven't pushed for the network upgrade.

I hope to figure out when, where and how to use generators and
decorators at some point, but I just haven't gotten that far a long
yet, I guess.

Mike

 

[Thomas Bellman]

Desktop or server?

If server, check what the major Linux distros, like Fedora
Core, are shipping with.

For server, you should probably rather look at distros like
RHEL/CentOS, Suse and Debian Stable.

For what it's worth, Fedora 8 has Python 2.5, RHEL 5 ships with
Python 2.4, and RHEL 4 has Python 2.3. Suse and Debian, I don't
know.

Check major shared hosting providers to see what they're offering
to their customers as standard.

I would expect that to often depend on what OS version they are
using. And RHEL/CentOS 4 is still quite common, so if you want
to reach a large "customer base", make sure that your Python
programs work with Python 2.3.

 

[Joseph Turian]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

 

[John Nagle]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

So I guess the question is, does anyone have a sense of what percent
of python users don't have 2.5?

Desktop or server?

If server, check what the major Linux distros, like Fedora
Core, are shipping with.

Check major shared hosting providers to see what they're offering
to their customers as standard.

John Nagle

 

[George Sakkis]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

So I guess the question is, does anyone have a sense of what percent
of python users don't have 2.5?

Perhaps you should ask the inverse question too: what 2.5 features do
you find so compelling that you are willing to break compatibility
with 2.4 ? FWIW, the only new 2.5 feature I have been using in
practice is the conditional expressions, and I could easily live
without them. 2.4 is still pretty decent, and a major upgrade from
2.3.

George

 

[Gabriel Ibanez]

Debian Etch (stable) has Python 2.4


----- Original Message -----
From: "Thomas Bellman" <[email protected]>
Newsgroups: comp.lang.python
To: <[email protected]>
Sent: Friday, April 18, 2008 8:50 PM
Subject: Re: Python 2.5 adoption

 

[Carl Banks]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

So I guess the question is, does anyone have a sense of what percent
of python users don't have 2.5?

One possible barometer for the situation is what's the oldest version
of Python to have been supported in the most bug-fix releases?

....In which case you need to maintain backwards compatibility with
2.3.

(I bring this up to illustrate that if there are people clamoring for
a 2.3 updates, there are probably quite a few supporting 2.4 as well.)


Carl Banks

 

[Graham Breed]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

How about Java users? Jython was recently at 2.2 (still is for all I
know). I'm pleased they've got that far because I like to know that
my code can run under Java and I like generators.

My web host uses 1.5.2. That is painful.

If you're assuming your potential users already have 2.4 then the
chances are they'll have upgraded to 2.5 by the time you've finished
anyway.


Graham

 

[castironpi]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

99% is a big percent. My 1% doesn't like something.

 

[Donald 'Paddy' McCarthy]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

Then develop for 2.5 with an eye on what is to come this year in 2.6 with regard to already planned
deprecations.

- Paddy.

 

[Ray Cote]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

A few seconds after reading this, I read the announcement for pyspread.
Requirements? Python 2.5.

You might want to talk with the pyspread folks regarding their
decision to require 2.5.
http://pyspread.sourceforge.net

--Ray

--

Raymond Cote
Appropriate Solutions, Inc.
PO Box 458 ~ Peterborough, NH 03458-0458
Phone: 603.924.6079 ~ Fax: 603.924.8668
rgacote(at)AppropriateSolutions.com
www.AppropriateSolutions.com

 

[Lie]

How widely adopted is python 2.5?

We are doing some development, and have a choice to make:
a) Use all the 2.5 features we want.
b) Maintain backwards compatability with 2.4.

There is another choice: Develop with future in mind because it's
possible that when you finished version 1, what seemed to be future
would become the present. This is especially if the project is big and
requires years to complete.

My web host uses 1.5.2. That is painful.

If them using 1.5.2 is painful for you ask the host to install
something newer (AFAIK it is possible to run several python versions
side-by-side) or prepare yourself to move to other host.

 

[Aahz]

Basically, we're planning on releasing it as open-source, and don't
want to alienate a large percentage of potential users.

Datapoint: my company still uses 2.3 and *might* upgrade to 2.4 and
later this year. Basically, any company with lots of servers has a good
chance to still be stuck with 2.2/2.3 (we only dropped 2.2 last fall).

 

[castironpi]

Datapoint: my company still uses 2.3 and *might* upgrade to 2.4 and
later this year.  Basically, any company with lots of servers has a good
chance to still be stuck with 2.2/2.3 (we only dropped 2.2 last fall).

Different is a verbally atomic relation.

 

[Lou Pecora]

Different is a verbally atomic relation.

It's a Passover question.

 

[castironpi]

It's a Passover question.

Did it Pass.

 

[Jorgen Grahn]

Perhaps you should ask the inverse question too: what 2.5 features do
you find so compelling that you are willing to break compatibility
with 2.4 ? FWIW, the only new 2.5 feature I have been using in
practice is the conditional expressions, and I could easily live
without them. 2.4 is still pretty decent, and a major upgrade from
2.3.

Another data point: I write some Python code in my work and some for
hobby/private use, and I am very happy with 2.3. List comprehensions
(or whatever they are called) and generators are the most recent
features I would hate living without.

OP: keep in mind that your users do not see any gain from you using
2.5. All they see is something that makes your software harder to
install. At some point you can dismiss them as living in the Stone Age,
but the Stone Age is currently 2.1 or something. Maybe 2.2 is, too.

/Jorgen

 

[Patrick Mullen]

OP: keep in mind that your users do not see any gain from you using
2.5. All they see is something that makes your software harder to
install. At some point you can dismiss them as living in the Stone Age,
but the Stone Age is currently 2.1 or something. Maybe 2.2 is, too.

Except for the memory bug which 2.5 fixed (not giving memory back),
which in some cases with 2.4 could be a really large issue. But in
that case you get the benefit whether it was "coded for" 2.5 or not.
So it is still safest to develop for a lower common denominator. For
me, the best things about 2.5 were the memory fixes/performance, and
the inclusion of ctypes in the standard library. The language
upgrades are mostly corner cases. If I were the OP, and those corner
situations aren't too big of an issue, I would restrict my usage to
2.4 or even 2.3 to allow easier adoption of the software.