PEP 236: Back to the __future__

Tim Peters tim.one at home.com
Tue Feb 27 12:13:15 EST 2001


[Karsten Petersen]
> ...
> I am not too lucky with all those improvements/changes inside the
> language.  In fact I am rather scared.
>
> Let me explain why:
> Since 1997 I am using Python for a lot of my projects, the reasons were
> ease of use, the clearness of the language and its availability and
> compatibility across platform borders.

I've been using it since 1991, and have found the burden of incompatible
Python changes to be trivial over that time.  They'll be even easier to deal
with using the __future__ mechanism.

> And well, I still use Python 1.5.2 and am not seeing a reason to
> upgrade.

That's OK.  The PEP doesn't *require* you to upgrade <wink>.

> If I upgrade to 2.0, I will have to recompile all C modules, check
> everything if it still runs and so on.  But what about 2.1 or 2.2?
> Will I have to recompile and/or recheck everything again?

Depends on the platform (e.g., on Windows the name of the release is in
effect buried in .pyds that link to it), and on whether changes are also made
to the Python virtual machine and/or C API.

> For me it seems so, and I am not willing to do that.
>
> I want to _use_ Python, its implementation details are something I am
> not really interested in.
>
> It is possible to compile C sources that were written in the 80s.
>
> It is possible to run (more or less simple) Perl programs which are
> several years old.
>
> Will I be able in 2005 to run the code I have written today?

I've got Python that hasn't changed since 1991; I expect it will work fine in
2011 too.  Don't overblow this.

> If there would be one big change I could cope with it.  But with smaller
> changes and trouble every couple of months I am rather unsatisfied.

No offense, but I don't buy this.  If you're still sticking to 1.5.2 by
choice, you're simply change-aversive.  Nothing wrong with that!  But it
makes the notion that some other form of change *would* leave you happy is
hard to buy.

the-pep-is-as-it-says-for-people-who-do-stay-current-ly y'rs  - tim





More information about the Python-list mailing list