Hacker News Reader

Comment by snprbob86 | Hacker News Reader

Is this more immediately obvious? I would have expected list to be defined as list(*elements) and called like this list(1, 2, 3)

I guess that this allows me to type help(list) and figure it out, but I would have done that anyway to identify an operator.

Clearly the answer is a .clone() or .copy() method...

View on HN

I think doing list(foo) to create a list copy is a pretty obvious way of getting around the 'variables are just references' thing.

I honestly have never seen someone do foo[:] just to copy a list, this is all news to me.

> Clearly the answer is a .clone() or .copy() method...

if you want Python to smell like Java or Ruby. :->

http://docs.python.org/library/copy.html

Seems to disagree with you.

That's a module in the standard library.

What I'm saying is, there's no copy or clone or dup method built into Python's base object, as there is in Ruby. And Python programmers are not instructed to define a clone method for a class, as in Java. A __copy__ or __deepcopy__, okay, but not copy or deepcopy.

Hence, in Python you just don't see foo.clone() or foo.dup() or foo.copy() or foo.deepcopy(). Instead, you see copy.copy(foo) or copy.deepcopy(foo). Pythonic code often looks different than Ruby or Java, looks more stand-alone function-y than instance method-y.

By putting it in a module, you avoid having to add the method to every single class.

Hmm, no. You have to add __copy__ and/or __deepcopy__ to every class.

Only to override the default implementation.

By putting it on Object (this is my ruby side speaking) you avoid the same thing.

Reply on news.ycombinator.com