General Q&A regarding Python 3, adoption etc.
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past several days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code. The two main additions I currently have planned are a question specifically about the state of the WSGI protocol (it works, but it an error prone way), as well as one on what I'd like to see as the next steps in encouraging Python 3 adoption now that we're within 18 months of the planned date for 2.7 to enter security fix only mode (which involve encouraging community workshops to switch to teaching Python 3.4 initially, with Python 2.7 as an optional follow up, helping Ubuntu & Fedora with their transitions to Py3 by default, bringing 3.5 closer to parity with Python 2 for wire protocol development, and, on the Red Hat/Fedora side, helping to encourage the adoption of software collections as a mechanism for decoupling the runtime for Python applications from the system Python on RHEL 6 and its derivatives. I thought I mentioned it on this list last year when I first wrote it, but some messages I've seen recently suggest many folks haven't seen it before. Cheers, Nick.
On 06/01/2014 23:16, Nick Coghlan wrote:
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past several days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code.
The two main additions I currently have planned are a question specifically about the state of the WSGI protocol (it works, but it an error prone way), as well as one on what I'd like to see as the next steps in encouraging Python 3 adoption now that we're within 18 months of the planned date for 2.7 to enter security fix only mode (which involve encouraging community workshops to switch to teaching Python 3.4 initially, with Python 2.7 as an optional follow up, helping Ubuntu & Fedora with their transitions to Py3 by default, bringing 3.5 closer to parity with Python 2 for wire protocol development, and, on the Red Hat/Fedora side, helping to encourage the adoption of software collections as a mechanism for decoupling the runtime for Python applications from the system Python on RHEL 6 and its derivatives.
I thought I mentioned it on this list last year when I first wrote it, but some messages I've seen recently suggest many folks haven't seen it before.
Cheers, Nick.
Is it on the back of a fag packet or is there a link somewhere? :) -- My fellow Pythonistas, ask not what our language can do for you, ask what you can do for our language. Mark Lawrence
On Mon, 06 Jan 2014 23:25:12 +0000 Mark Lawrence <breamoreboy@yahoo.co.uk> wrote:
On 06/01/2014 23:16, Nick Coghlan wrote:
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past several days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code.
The two main additions I currently have planned are a question specifically about the state of the WSGI protocol (it works, but it an error prone way), as well as one on what I'd like to see as the next steps in encouraging Python 3 adoption now that we're within 18 months of the planned date for 2.7 to enter security fix only mode (which involve encouraging community workshops to switch to teaching Python 3.4 initially, with Python 2.7 as an optional follow up, helping Ubuntu & Fedora with their transitions to Py3 by default, bringing 3.5 closer to parity with Python 2 for wire protocol development, and, on the Red Hat/Fedora side, helping to encourage the adoption of software collections as a mechanism for decoupling the runtime for Python applications from the system Python on RHEL 6 and its derivatives.
I thought I mentioned it on this list last year when I first wrote it, but some messages I've seen recently suggest many folks haven't seen it before.
Cheers, Nick.
Is it on the back of a fag packet or is there a link somewhere? :)
I assume it's http://python-notes.curiousefficiency.org/en/latest/python3/questions_and_an... Regards Antoine.
On 7 Jan 2014 07:46, "Antoine Pitrou" <solipsis@pitrou.net> wrote:
I assume it's
http://python-notes.curiousefficiency.org/en/latest/python3/questions_and_an... Thanks, that's the one - I copied the link, but neglected to paste it in before hitting send :P Cheers, Nick.
Regards
Antoine.
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe:
https://mail.python.org/mailman/options/python-dev/ncoghlan%40gmail.com
On Mon, Jan 6, 2014, at 03:16 PM, Nick Coghlan wrote:
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past several days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code.
I'd like to thank you for taking on the task of Python 3 justification.
On Tue, 7 Jan 2014 09:16:10 +1000 Nick Coghlan <ncoghlan@gmail.com> wrote:
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past several days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code.
A couple remarks: - the unicode section would gain being a little more on the practical side; for example the "surrogateescape" paragraph is an obscure and theoretical way of saying unicode filepaths (etc.) are fully supported on all platforms - also, it doesn't seem very clear that the primary string type (str) is now unicode; this has important consequences, for example non-ASCII exception messages work fine in 3.x while they were very delicate to work with in 2.x - when discussing Twisted / gevent alternatives, you should also mention Tornado, which is especially interesting because it works on both Python 2 and Python 3, and therefore presents a nice migration path - perhaps you should discuss the idea that "uptake is slow", because the numbers are rather conflicting on that point; see what I wrote in https://mail.python.org/pipermail/python-list/2014-January/663922.html and also Chris Angelico's elaboration in https://mail.python.org/pipermail/python-list/2014-January/664003.html Regards Antoine.
On 7 Jan 2014 08:03, "Antoine Pitrou" <solipsis@pitrou.net> wrote:
On Tue, 7 Jan 2014 09:16:10 +1000 Nick Coghlan <ncoghlan@gmail.com> wrote:
For anyone that isn't already aware, I wrote a Q & A about Python 3 last year (in response to an article about how we should have fixed the GIL instead of Unicode), and I've updated it extensively over the past
several
days due to Alex's misunderstanding of the objectives for Python 3.4 as well as Armin's latest piece on the increased difficulties in writing wire protocol handling code.
A couple remarks:
- the unicode section would gain being a little more on the practical side; for example the "surrogateescape" paragraph is an obscure and theoretical way of saying unicode filepaths (etc.) are fully supported on all platforms
- also, it doesn't seem very clear that the primary string type (str) is now unicode; this has important consequences, for example non-ASCII exception messages work fine in 3.x while they were very delicate to work with in 2.x
- when discussing Twisted / gevent alternatives, you should also mention Tornado, which is especially interesting because it works on both Python 2 and Python 3, and therefore presents a nice migration path
Thanks, I've addressed these and a couple of other points people brought up (e.g. it is cx-freeze that supports Py3k, not py2exe).
- perhaps you should discuss the idea that "uptake is slow", because the numbers are rather conflicting on that point; see what I wrote in https://mail.python.org/pipermail/python-list/2014-January/663922.html and also Chris Angelico's elaboration in https://mail.python.org/pipermail/python-list/2014-January/664003.html
I haven't incorporated these observations yet, but I will. It ties in closely with the point that bootstrapping the new Python 3 application ecosystem with cross-version libraries and frameworks is not the same thing as migrating the existing Python 2 *application* ecosystem, and the latter is expected to take *much* longer (since existing Python 2 users will have, of necessity, already worked around or avoided the bugs and limitations of that version of the language). Cheers, Nick.
Regards
Antoine.
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe:
https://mail.python.org/mailman/options/python-dev/ncoghlan%40gmail.com
participants (5)
-
Antoine Pitrou -
Benjamin Peterson -
Emile van Sebille -
Mark Lawrence -
Nick Coghlan