The potential for a Python 2.8.

[Chris Angelico]

On Fri, Jan 24, 2014 at 11:42 AM, Terry Reedy <tjreedy at udel.edu> wrote:
> Burying 'Python 2.8' was the purpose of PEP 404. It is kind of bizarre.
> Developers informally said 'No 2.8'. People would not believe that. So
> developers formally said 'No 2.8'. They even inverted the purpose of PEP to
> make the formal announcement visible and permanent. And a few people still
> do not want to believe it.

Can I get a new version of Java 1.1.8 please? I want it to include all
the cool features that I want from the newer versions, but it has to
still run all my existing code. I'm not going to put in any effort to
actually _make_ this, I want you to do it for me.

Actually, the Java versioning system was enough of a mess that, to
this day, I don't know what version(s) my old Java code would and
wouldn't run on. So glad to have moved away from that. At least with
Python, semantic versioning [1] means everyone knows what everyone's
talking about. Python 2.8 has to be broadly compatible with 2.7 and
doesn't have to be compatible with 3.3. (Which, incidentally, is at
odds with some people's idea of a 2.8, which would be incompatible
with both. I'm not sure what that would be called - e.1? sqrt(8).0?
Something else?)

The noise asking for a 2.8 isn't going to die down any time soon.
It'll flare up again every time there's a significant event in the
2.7's end of life: when it goes into source-only support, when its
python.org support ends entirely, when Debian's next version won't
ship it, when Red Hat's ditto ditto, when it's no longer possible to
get it from Ubuntu's repositories, etc, etc, etc. And no amount of
"There will be no 2.8 unless you make it yourself!" will change that.

That's my prediction.

ChrisA

[1] Not sure if Python's actually stated that http://semver.org/
principles are guaranteed to be followed, but they're certainly
approximated to