evaluation question

[avi.e.gross at gmail.com]

I think its has been discussed here that many functions are DELIBERATELY
designed to return without returning anything. Earlier languages like Pascal
had explicit ideas that a function that did not return a value was declared
as a "procedure" but many other languages like python make no real
differentiation.

Some functions are designed for a sort of side-effect and often there is
nothing much that needs to be returned or even can be. If a function prints
a dozen items one at a time, should it return nothing, or a copy of the last
item or somehow of all items? Generally nothing looks right. If you want to
return something, fine. Do it explicitly.

Similar arguments have been made about methods that do things like sort the
contents of an object internally and then return nothing. Some would like
the return to be the (now altered) object itself. You can emulate that by
not sorting internally but instead sorted(object) returns a new object that
has been sorted from the old one.

So should or could print return anything? Other languages exist, like R,
that do return (and often ignore) whatever print displayed elsewhere. This
can be of use in many ways such as making it easier to print or store
additional copies without recalculating. 

My preference might be to simply allow a local option at the end of a print
statement such as print(..., return=True) or even a way to set a global
option so all print statements can be turned on when you want. But is this
pythonic? In particular, people who want to give type hints now can safely
claim it returns None and would have to modify that so it can optionally
return something like str or None. And, of course, once you change print()
this way, someone else will want the number of characters (or perhaps bytes)
returned instead.

Much of this can be worked around by simply making your own customized print
function which evaluates the arguments to make a string and then calls
print, perhaps with the results pre-calculated, and returns what you wanted.
That is not as easy as it sounds, though as print  supports various
arguments like sep= and end= and file= and flush= so a weird but doable idea
is simply to substitute a temporary file for any file= argument and write
the results to a temporary file or something in memory that emulates a file.
You can then read that back in and return what you want after handling the
original print statement with the original arguments, or perhaps just use
your result to any actually specified file or the default.

You can thus create something like what you want and leave the original
print() command alone to do what it was designed to do.

And, in general, people who want a copy of what they print, often use other
python functionality to craft some or all parts of the text they want
printed and only then call print() and thus already may have the ability to
use the text afterwards.

For many purposes, including efficiency, returning nothing makes good sense.
But it is not really the only choice or the right choice and yet, if you
want to use THIS language, it has to be accepted as the documented choice.


-----Original Message-----
From: Python-list <python-list-bounces+avi.e.gross=gmail.com at python.org> On
Behalf Of Thomas Passin
Sent: Tuesday, January 31, 2023 1:16 PM
To: python-list at python.org
Subject: Re: evaluation question

On 1/31/2023 4:24 AM, Muttley at dastardlyhq.com wrote:
> On Tue, 31 Jan 2023 12:57:33 +1300
> Greg Ewing <greg.ewing at canterbury.ac.nz> wrote:
>> On 30/01/23 10:41 pm, Muttley at dastardlyhq.com wrote:
>>> What was the point of the upheaval of converting the print command 
>>> in python 2 into a function in python 3 if as a function
>>> print() doesn't return anything useful?
>>
>> It was made a function because there's no good reason for it to have 
>> special syntax in the language.
> 
> All languages have their ugly corners due to initial design mistakes 
> and/or constraints. Eg: java with the special behaviour of its string 
> class, C++ with "=0" pure virtual declaration. But they don't dump 
> them and make all old code suddenly cease to execute.
> 
> Pragmatism should always come before language purity.
> 

It was more fundamental than that, and not mainly about print():

https://snarky.ca/why-python-3-exists/
--
https://mail.python.org/mailman/listinfo/python-list