Noob in Python. Problem with fairly simple test case

Rick Johnson rantingrickjohnson at gmail.com
Fri Jul 17 02:30:33 CEST 2015


On Thursday, July 16, 2015 at 6:24:21 PM UTC-5, Chris Angelico wrote:

Any attempt to translate downloads into *REAL* usage
statistics is doomed to be unreliable. Chris, you're smarter
than this!

(1) for instance: Python2.x coders have been around long
enough that they don't need to download as much from PyPi
anymore. Whereas, Python3.x coders are more recent, therfor
they will be downloading the majority of packages these
days.

(2) Also. Old and stable code bases do not change as much as
new unstable code bases tend to do. Therfor, less Python2.x
downloads

(3) The numbers are falsified even more by 3.0 packages that
are downloaded, but never actually used. Maybe the package
was not what the downloaded expected it to be. I have
downloaded many packages from PyPi and never used the
majority of them.

> 
> Even if you don't correct for that, these figures show *at
> most* 30-50% more Py2 usage than Py3, which hardly
> justifies Rick's statement that the "vast majority" of
> Python is 2.x.
> 

I think I've dispelled the validity of your assertion and
the weakness inherent in your mental heuristic. As my
logical reasoning demonstrates, your percentages of 30%-50%
are, at best, on the *LOW* end of the real percentages of
Python2.x usage.

And i'm not talking about downloads. I'm talking about
*REAL* code out in the wild. I talking about real
applications out in the wild, or living on hard drives doing
*REAL* work. That's the *REAL* litmus test of Python!

Not some false download statistics that can easily be
tainted by a large number of 2.x coders downloading and
"playing" with 3.0 source and packages, but never releasing
or even creating anything substantial. 

Where is the substantial amount of Python3000 applications,
code, and libraries? Sure you can point to PyPi, as it does
contain 3.0 compatible code, but that code is sitting around
doing nothing?

But more importantly than *WHERE* -> *WHAT* are they doing?

Are they making peoples lives better? Are they solving
problems? Or are they just toys written to satisfy some
childish need to play, before getting bored and moving on to
real life projects?

The fact is, the *REAL* code that is doing *REAL* work that
is solving *REAL* problems, is doing it in Python2.x That's
a fact Chris. And your smoke and mirror parlor tricks are
not going to change that one bit.




More information about the Python-list mailing list