[Python-ideas] Statement local functions and classes (aka PEP 3150 is dead, say 'Hi!' to PEP 403)

[Georg Brandl]

Am 17.10.2011 02:30, schrieb Steven D'Aprano:
> Nick Coghlan wrote:
>> On Sun, Oct 16, 2011 at 9:01 PM, Steven D'Aprano <steve at pearwood.info> wrote:
>>> If it's not tested, how do you know it does what you need?
>> 
>> Because my script does what I want. Not every Python script is a
>> production application. We lose sight of that point at our peril.
>> 
>> Systems administrators, lawyers, financial analysts, mathematicians,
>> scientists... they're not writing continuously integrated production
>> code, they're not writing reusable modules, they're writing scripts to
>> *get things done*.
> 
> "And it doesn't matter whether it's done correctly, so long as SOMETHING 
> gets done!!!" <wink>

Sorry, but even with the wink this is a dangerous statement.  When I write
a 40-line script to read, fit and plot a dataset, I don't add a unit test
for that.  On the contrary, I want to write that script as conveniently
as possible.  Python is as popular as it is in science *because* it allows
that.  People hate writing boilerplate just to do a small job.

> I'm not convinced that requiring coders to write:
>
> given a, b c
> do f(a, b, c)
>
> instead of
>
> do f(a, b, c)
> given a, b, c
>
> gets in the way of getting things done.

I guess this means -0 from my side given:

   That said: I'm not sure this specific proposal would help.  I agree with
   you and can't see that defining the function before using it is the wrong
   order.  What does appeal to me is the potential for cleaner namespaces;
   while this doesn't matter so much in the function-locals (there it's
   mostly the OCD speaking), it would definitely helpful for module and
   class namespaces (where we routinely see "del" statements used to remove
   temporary names).

   I can believe that others see the "given" semantics as the natural order
   of things.  This is probably why Haskell has both "let x = y in z" and
   "z where x = y" constructs to locally bind names, but there it fits well
   with the syntax, while in the case of Python, the "given: suite" feels
   somewhat out of place; we really have to think hard before adding new
   syntax -- one more thing that the "casual" users Nick mentioned have to
   grok.


cheers,
Georg