[issue39416] Document default numeric string formats

[Karl O. Pinc]

Karl O. Pinc <kop at karlpinc.com> added the comment:

On Wed, 22 Jan 2020 06:09:41 +0000
"Eric V. Smith" <report at bugs.python.org> wrote:

> Eric V. Smith <eric at trueblade.com> added the comment:
> 
> Is the lack of this documentation causing some confusion somewhere?
> This isn't rhetorical, I'm genuinely curious what problem you're
> trying to solve.

I'd say no, there's no real confusion anywhere.

This started with a search for documentation on Python's
default string representation for float.  This accidentally
wound up on the python-dev mailing list and became a
more general discussion during which I suggested that
some fundamentals are clear.  This is pretty much just
a follow-up to that discussion.

> Is there any mainstream programming language where the basics of what
> you've laid out aren't true? It all seems pretty obvious to me. For
> example, I can't see anyone looking at this and saying "Ah, base 10.
> That's why it's producing the output I'm seeing."

Well, I just got done looking at YAML which has a lot of base-related
syntax including base 60.  :)

The problem that would be solved is that it's easy to rely
on the default Numeric output formats.  Most output probably
does not go through a formatter, although this may change
now that f-strings are so easy.  Anyway, this would guarantee
"normal output" for numbers, even when switching between
Python implementations.

Likewise, immutability of value when round-tripping through a
string is also obvious.  But is still an important property.

So if these properties are important then include them in the spec.
That's what specs are for.  (I could probably find some non-mainstream
languages (scheme perhaps) that have specs which include
similar documentation.  You pretty have to have a formatter
in compiled languages, so those won't specify a default.
Anyway, it shouldn't matter what other language specs do.)

I understand if nobody sees this as a real problem.  And
I don't want to stand up as a big proponent.  I'll
point out the argument here and let others decide.

Regards,

Karl <kop at karlpinc.com>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein

----------

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue39416>
_______________________________________