How do you explain the fact TensorFlow targets Python 3 for speed, features, etc.?
Most of their official examples do not work in Python 3. https://github.com/tensorflow/models
Google will kill Python 3, and it might be a good thing
21–25 of 25 posts
Re: Google will kill Python 3, and it might be a good thing
#22I think the choice of title for HN is unfortunate: it implies (for me) some inside knowledge about what is (or will) happen to Python. Casual HN readers will see the death of Python as a done deal, whereas the truth is "the reports of [Python's death] are much exaggerated".
Re: Google will kill Python 3, and it might be a good thing
#23i totally disagree with your view. Python 3 of today is fantastic and the community is fully behind it. Yes there will be legacy code that will remain on python2 for ever, but this is normal. I am personally loving the new features of python 3.6 in particular and would not want to look back. Google has a lot of code in python2 they want to move to go. It would be bizarre to see anyone start a new project in python 2…
With due respect, if the community was fully behind Python 3, it wouldn't still be controversial. Clearly it still is.
That said, my personal feeling is that with awesome new toys like curio, Python 3 is finally starting to be worth the pain of cutting over.
Re: Google will kill Python 3, and it might be a good thing
#24i totally disagree with your view. Python 3 of today is fantastic and the community is fully behind it. Yes there will be legacy code that will remain on python2 for ever, but this is normal. I am personally loving the new features of python 3.6 in particular and would not want to look back. Google has a lot of code in python2 they want to move to go. It would be bizarre to see anyone start a new project in python 2…
With due respect, if the community was fully behind Python 3, it wouldn't still be controversial. Clearly it still is. That said, my personal feeling is that with awesome new toys like curio, Python 3 is finally starting to be worth the pain of cutting over.
> With due respect, if the community was fully behind Python 3, it wouldn't still be controversial. Clearly it still is.
Personally, as of PyCon last year, I do not think it's controversial anymore. If anything still doesn't support Python 3, it's at the service, such as AWS Lambda. The libraries have all transitioned by now [1].
Re: Google will kill Python 3, and it might be a good thing
#25I think the argument for runtime fragmentation and package fragmentation is a bit exaggerated. Most people use pip. Most people use CPython. People make tradeoffs when they need to use another interpreter for performance reasons [1], but CPython is the reference implementation. Just because alternative options exist for specific cases doesn't mean there's fragmentation in the ecosystem at large.
[1]: http://docs.python-guide.org/en/latest/starting/which-python...