Migrating to python3

I was recently asked, by my boss, about the possible benefits of migrating our code base at work from python 2.6 to python 3. Instead of sending an email back to my boss with my answer, I decided that it worth a blog post. Why? because the internet is full with comparisons between python 2 and python 3, but it is hard to tell what are the real gains from doing the migration. Here, I will try to explain, in a very subjective manner, the benefits I see from such move.

So let's start with the obvious, python 3 is clearly a better language than python 2.6: print is a function instead of a statement, now with sane arguments; dict and set comprehensions reduce unnecessary, error prone, verbosity; real division by default, because that's what you want 99% of the time anyway; iterators by default (range replaced xrange, dict.items() replaced dict.iteritems(), etc.); explicit relative imports... Actually, the list goes on and on. However, I don't think that the features above are good reasons to migrate a large project. Modern python code, written with python 3 in mind, will benefit from them, that's for sure. But we do not want to rewrite our code base, we want the transition to be as smooth as possible.

So what is valuable? Here is my list:

>>> # python2
>>> 'hello' < 1

>>> # python3
>>> 'hello' < 1
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unorderable types: str() < int()

>>> # python2
>>> i = 42
>>> squares = [i ** 2 for i in xrange(10)]
>>> print(i)
9

>>> # python3
>>> i = 42
>>> squares = [i ** 2 for i in xrange(10)]
>>> print(i)
42