That’s because Google uses Python 2.7 internally. They have a ton of code for it, and any migration is harder than they might seem.

As per public statements, Google has ~100 million lines of Python code. Also, Google has used Python very heavily from the start, meaning that some of that code was written in Python 1.x or early 2.x days, before the language had packages, before standard modules like 'logging' or 'unittest', even before booleans(!).

It's just a lot of code to migrate. It takes time.

Disclaimer: Work at Google, wrote some of that code.

Sure, but why not support it in an external facing product? No one externally depends on that old google code.

Started as an internal tool, at least according to medium & other blog posts.

It is in the works [0], and like people said since Google uses Python 2.7 that's the verison that was supported by default.

[0] https://stackoverflow.com/questions/46941308/python-3-suppor...

They can release a version that supports 2.7 much sooner

Jupyter switch really easily from Python2.7 to Python 3 to C# because the front end and backend are separated. Not sure why they are having issue here.

What's wrong with Python 2.7? (real question, not flamebait)

Python 2.7 is frozen. It just get security updates. The community is investing in Python 3. All the main numerical libraries are already compatible with Python 3 and, unless you need a very specific library, there is no reason to start developing new projects in Python 2 nowadays.

Unicode text handling.

This is wrong with Python 3 as well in real world use.

How so?

Maybe the parent commenter is referring to issues brought up by Armin Ronacher (among others): http://lucumr.pocoo.org/2014/5/12/everything-about-unicode/

Don't know what has changed since 2014. My perception is that Armin feels he's been pigeonholed as being more anti-Py3 than he actually is, at least with the state of Python in 2016:

http://lucumr.pocoo.org/2016/11/5/be-careful-about-what-you-...

> In fact, I myself campaigned for some changes to Python 3 that made it possible to achieve better ports (like the reintroduction of the u prefix on Unicode string literals) and the bulk of my libraries work on Python 3 for many years now. It's a fact that in 2016 the problems that people have with Python 3 are different than they used to have before.

See similar thread made yesterday: https://news.ycombinator.com/item?id=15647727

That didn't say anything about what's wrong with 2.7