Why is Python 3 (or later) better than Python 2?

I learned Python as my first serious (non BASIC) language about 10 years ago. Since then, I have learned lots of others, but I tend to 'think' in Python. When I look at the list of changes I do not see one I need this feature. I usually say to myself, hmm that would been a good way of doing it, but why change it now?

Things like changing the default floor division could be a real pain to change for big projects. It seems like the major players are dragging their feet. What开发者_开发技巧 is the key feature that would make me want to invest in another learning curve?

---

As a key feature, a lot of people seem to be pretty exited about (supposedly) transparent unicode support. They changed it from str (8-bit char array/default string type) and unicode (unicode string), to str (default (unicode compatable) string) and bytes (binary data as 8-bit 'string').

(I think seperation of byte lists from strings is great idea, but I also hate unicode, so if anything, this would be a worse for me personally.)

---

A good discussion of this can be found in the python wiki; Should I use Python 2 or Python 3 for my development activity?

---

Things like changing default floor division could be a real pain to change for big projects.

If you had started making the change 8 years ago when Python 2.2 was introduced with // and from __future__ import division, it wouldn't be a pain now. Personally, I'm glad to finally get rid of old-style division!

My second-favorite feature of Python 3.x is the str/bytes distinction. Besides making Unicode support easier, bytes is far more convenient for database BLOBs than buffer was.

---

On Teaching Programming With Python 3.0, though a bit dated, is one of the best articles I've read on the advantages of Py3k.