PEP0238 lament

[Chris Gonnerman]

----- Original Message ----- 
From: "Greg Ewing" <greg at cosc.canterbury.ac.nz>


> Arthur_Siegel at rsmi.com wrote:
> > 
> > If I was told that experinced programmers might
> > appreciate this change for any possible reason -
> > I'd be quiet as a mouse.
> 
> I'm an experienced programmer, and I appreciate the
> change on the grounds that different operations on
> the same data should be spelled differently.
> 
> You may lapse into silence now. :-)

Well, I'm an experienced programmer also, and I don't 
appreciate the change.  Ultimately that's the entire
reason... code written for 1.5.2 still works great in 
2.1, and will work fine in 2.2 and apparently 2.3, but
will then fail in 2.4.

Suppose that I follow all this, dutifully finding and
fixing all the warnings in my modules as we move from
version to version; this is the plan, after all.

Now, here I am some years later and I look for a module
to do something or other, and on the Vaults I find one; 
but the module (which may be huge) was written for 1.5.2.
I have to debug someone else's code (a big job even in
Python), find another source (if there is one) or write
the functionality myself.

If we don't screw with the integer division, I'm probably
in good shape.  If we do, bang.

Changes that are mostly backwards compatible are cool, 
but when something as fundamental as division of integers
changes, it gets pretty ugly.