Quite frankly [LWN.net]

Quite frankly

Posted Feb 17, 2011 19:51 UTC (Thu) by NikLi (guest, #66938)
Parent article: Moving to Python 3

i don't see any new features in Python 3 that would make me want to switch.

I mean, "why?". And then there is a certain pressure to make the switch and show to the world that Python3 was a huge success, unlike the "perl6 fiasco", and then show the success of the python development model, PEPs, BDFL, etc. This seems like burning bridges as in "if we go back to python2, we will be like those perl6 people we've been making fun of".

The techinal differences are minimal and many of them in the domain of "nits".

I find arguments that say that "in the future there will be only Python3" misleading. There are many firms that use python extensively for their services and THEY WON'T SWITCH ANY TIME SOON. Count google as one of them, and i bet you they haven't got the least interest to rewrite their infrastructure to python3 because "dict.keys is an iterator", etc.

And, last but not least, YOU DON'T BREAK "Hello World".

Quite frankly in education i'd rather teach my students Python2, so they'll go work for google or something...

(Log in to post comments)

Quite frankly

Posted Feb 17, 2011 21:17 UTC (Thu) by foom (subscriber, #14868) [Link]

I attempted to ask the "are you really sure?" question a few different times during the development of 3.0, but obviously the answer was "yes!" every time.

But, at this point, there have been 3 releases of 3.x made now, and 2.x isn't being developed anymore. So, if you want any new features ever, you either have to pick up maintenance of 2.x, or switch to 3.x.

I wouldn't say it's impossible to conceive of the first alternative happening, but I'm certainly not interested in doing that, even though I'd really prefer if 3.x just magically ceased to exist. In another 2-3 years when 3.2 is installed ubiquitously alongside 2.x, maybe I'll even start writing python3 code. Stranger things could happen. :)

2.x features and improvements

Posted Feb 21, 2011 23:26 UTC (Mon) by pboddie (subscriber, #50784) [Link]

But, at this point, there have been 3 releases of 3.x made now, and 2.x isn't being developed anymore. So, if you want any new features ever, you either have to pick up maintenance of 2.x, or switch to 3.x.

However, most of the implementations apart from CPython work with 2.x features, and although there have been noises amongst some of them about 3.x support being a possibility, the priorities of their developers would appear to be the development of other kinds of features than language features. So, for example, PyPy would seem to be sticking with 2.x support and concentrating on performance - it's already faster than CPython for some things and getting faster - so if you value those kinds of features over the language tidying that 3.x represents, then you're not going to switch to 3.x.

But then again, neither are the implementation developers, so 2.x is still a very safe bet.