This Week in Programming: Goodnight, Sweet Python 2.x

The time has finally come to bid adieu to Python 2.7, which as of Jan. 1, 2020 is officially dead. Well, almost completely officially dead.
For those of you who have been watching, the Python Clock has counted its way down to zeros across the board after nearly five years on the job, having been created upon the adoption of a new end-of-life deadline for Python 2.7. Originally, the language was set to expire in 2015, but was given an extension “to clarify the status of Python 2.7 and relieve worries for those users who cannot yet migrate to Python 3.” Give a click through to that email thread and you’ll see that this day has been a long, long time in the coming.
Now, back to that officially versus unofficially dead part. Python 2.7 is officially dead, but we haven’t seen the last of it quite yet. According to a statement from the Python Software Foundation, the last major version 2.7 will be released in April 2020 (to coincide with PyCon), and then all development will cease for Python 2.x. The foundation further explains the five-year delay, writing that “the stricter text model in Python 3 was forcing the resolution of non-trivial Unicode handling issues in the reference interpreter and standard library, and in migrated libraries and applications” and that the extension was given to give people adequate time to migrate. So, yes, Python 2.x is dead from a further development standpoint, though we will still see one final release, and for those using systems like Red Hat Enterprise Linux, you have a little longer to go before the rug is officially pulled out from under your feet.
But, how is that migration to Python 3.x coming along, anyways? It sort of depends on who you ask.
JetBrains has run a Python developer survey for some years now and shows a continual decline in Python 2.7 use, with 25 percent still clinging to 2.7 in 2017, 16 percent in 2018, and just 13% in 2019. But perhaps those numbers are skewed a bit by the audience being surveyed, because a StackOverflow blog post from just two months ago takes a different approach and comes up with quite a different result. By contrast, StackOverflow asks why the migration to Python 3 is taking so long, and rather than asking people what version they’re using, they look at some numbers regarding downloads of PyPI, the Python package library. These numbers also show a decline, but tell a different story:
“Even though Python 3 has been the dominant version in the community for at least a year, the latest count of individual package downloads from PyPI shows that at least 40% of all packages for September of 2019 are 2.7 downloads. Granted, this is down from 60% at the beginning of the year, but is still significant given that EOL is only several months away.”
If you still haven’t made the leap, whether you’re the 13% (or 40%) of Python users, you might want to check out the official porting guide — lots of projects have pledged to drop Python 2.x support. Me, I’m just hoping that the official end of life might mean that I can give my Python tutorial another shot and maybe make it past the setup this time around.
Now, as for news in the world of programming, what with two mid-week holidays here, there’s little to be had. So, instead, we bring you a smattering of yearly round-ups and predictions for the year ahead. You’re welcome. To start, here’s a developer round-up of 2019 from Google:
This Week in Programming
- The Best APIs of 2019: Next up, there’s ProgrammableWeb’s yearly end-of-year series that looks at its “most clicked, shared and talked about APIs of 2019” which span a wide array of categories, including Advertising, Customer Relationship Management, Content, Printing, Products, Catalogs, Subscription, Feedback, Monetization, Auctions, Shipping, Barcodes, Ratings, Customization, Surveys, SEO, Tweets, and Engagement. There’s not much to explain there — each article takes a look at the top APIs, according to those metrics, for each category and provides a rather comprehensive look at the best and brightest of the 2019 API class. If that’s not enough for you though, the site also offers up its top 10 API articles of 2019 based on Twitter traffic, so take a gander.
Don't push to production on holidays.
You may end up sending an email to ~650 people with placeholder content instead of actual email content.
Not that I'm speaking from experience or anything. That would just be silly.
— Kelly Vaughn (@kvlly) January 1, 2020
- Predictions and Retrospectives: As for predictions for the year ahead, InfoQ offers a 2019 retrospective that touches upon many of the topics we talk about here — quantum computing, WebAssembly, Java, .NET, DevOps and so on — which is part of a larger collection of articles, videos, what have you, all on the topic of looking back at 2019 and ahead at 2020. SDTimes, meanwhile, jumps on the predictions train with software stability predictions for 2020, one of those lovely predictions pieces that gathers a bunch of pre-fabricated quotes most likely written by PR people, and takes a look at Gartner’s top ten trends for 2020. There’s even a look at what software will be written in 2020, as if it would be drastically different from 2019. Next up, PaktPub examines the 10 tech startups for 2020 that will help the world build more resilient, secure, and observable software, a list that includes some names you’ve also seen in The New Stack’s pages — LaunchDarkly, NS1, and more. And finally, as far as these sorts of things go, Amazon Web Services pats itself on the back with a review of its contributions to open source in 2019.
DockerCon is going virtual in 2020! The June 15-18, 2020 event in Austin, TX will no longer take place & will be a virtual event instead. Thank you for being a part of the DockerCon community! See you at the virtual event 🐋
— DockerCon (@DockerCon) December 30, 2019
- So You Wanna Learn in 2020? Crystal balls and nostalgia aside, we come to the other aspect of a new year – the resolution. We’ve all, unfortunately, heard the “New year, new me” thing by now, and developers and developers-to-be appear to be no exception. Everyone is out there pledging to learn a new language, and HackerNews seems to be the place for advice, as a couple of top threads this past week have hit on the topic of learning. So, if you plan on learning about machine learning or just getting clear on some computer science fundamentals, these threads are a good place to start.
If you're goal is to learn how to code in 2020, here's a thread of advice for you 👇🏻
— Ali Spittel 🐞 (@ASpittel) January 1, 2020
Amazon Web Services and NS1 are sponsors of The New Stack.
Feature image: “The Great Wave,” by Katsushika Hokusai (1830-32), New York Metropolitan Museum of Art.