PEP0238 lament

Stephen Horne steve at lurking.demon.co.uk
Mon Jul 23 23:34:06 CEST 2001


On Mon, 23 Jul 2001 19:01:56 GMT, "Tim Hochberg"
<tim.hochberg at ieee.org> wrote:

>It seems you would be better served proposing:
>
>def simple_func(x, y):
>   directive use_float_division
>   return y / x**2

I suggested the option to use...

def simple_func(float x, float y):
   return y / x**2

Where the types declarations are more like explicit coercions than
than types. The suggestion was that such implicit coercions should be
allowable for any identifier (attached to the indentifier, not the
value) and thus any assignment could then easily be checked.

>Maybe you already have and it was lost in the noise? However, I suspect that
>your more likely to get more mileage out of pushing for the ability to use:
>
>directive old_style_division

I suggested a direct equivalent for this, and an alternative that
simply turns on paranoid checks so that people could use a *new*
operator for the *new* functionality.



There has been a lot of noise - but there is a good reason for that
noise. It is quite clear that the pro-position is *not* an absolute
definitive fact - it is an opinion, same as the anti position and
definitely no stronger.

I'm not worried about Perl advocates laughing me, but I've been a very
vocal Python advocate at work and now it looks like a fundamental
arithmetic operator is going to change. I *AM* worried if my
professional reputation is going to be seriously damaged all because I
trusted in Python. THAT is a very serious issue that could serious
decrease the number of Python advocates. I for one am not going to
stick around if it's just to have myself made to look a complete arse.

Here's the pitch... "Hey, why not advocate Python? - professional
humiliation and career suicide can be yours today!"




More information about the Python-list mailing list