Ian Kelly wrote:
On Wed, Dec 14, 2011 at 1:28 AM, Felipe O <pip....@gmail.com> wrote:
Hi All,
I was wondering what everyone's thought process was regarding properties.
Lately I find I've been binging on them and have classes with > 10
properties. While pylint doesn't complain (yet), it tends to be picky about
keeping instance attribute counts low, so I figure there's something against
that. How do you guys decide between using properties versus getter methods,
or how do you refactor them if neither?
I prefer direct instance attribute access where possible*, properties
where necessary, and methods where an argument is needed or the
relationship is more complex than get/set/delete.
* One of the strengths of Python's property system** is that you can
switch between plain attributes and mutable properties as needed
without breaking dependent code. Often I see people doing this, which
drives me nuts with its useless verbosity, when a plain instance
attribute would have sufficed:
@property
def foo(self):
return self._foo
@foo.setter
def foo(self, value):
self._foo = value
** As opposed, for instance, to the .NET property system. You can't
arbitrarily switch between public member variables and public
properties in .NET, because it breaks ABI.
Cheers,
Ian
I second this opinion, plain attributes are what's required most of the
time.
JM
--
http://mail.python.org/mailman/listinfo/python-list
