Copy constructors

Alex Shindich alex at shindich.com
Tue Aug 14 22:32:40 EDT 2001


As always, Mr. Martelli’s posts are long and full of fascinating
detail. This time though, I have managed to read it all the way
through without forgetting the beginning of the post. In fact, I am
fully and entirely in support of Alex’s position! I do care for
the new features in PEP 252, but I think that the default behavior
should not change. Here are my reasons:
1.	I do believe that "premature" optimizations are evil. They usually
lead to really clumsy design decisions with very little performance
gain.
2.	I am worried about incompatibility with older versions of Python.
As a person who does have some amount of production code written in
Python, I am worried about the effort of making my code compatible
with this change. Even though I have unit tests for most of my code,
it is still an effort to fix all the places where I need to turn on
"dynamism".
3.	Most Python developers did not choose Python over C++ for speed.
Python is attractive because it is a very clean OO language with
dynamic behavior. Making Python less dynamic and “agedly”
faster will not attract more developers. Making it less dynamic will
be a turn off for some members of the existing Python community.

Just thought I would offer my 2 cents on this issue.



More information about the Python-list mailing list