Pointless Fanfare around Python Deprecating Unused Standard Librariesh

Published

About a week ago an article popped up around the internet called "Python: Batteries Included, But They're Leaking." The article talks about all of the problem with having a large standard library, including difficulty maintaining it.

It then goes on to make some pretty far reaching statements like it "crowds out innovation," stating that people won't use a potentially innovative third party library if something in the standard library exists that solves the problem. Has this ever been true? People still use rapidjson or simplejson for performance despite the fact that Python has a built in (and great) json library. Requests and urllib3 are as popular as they've ever been. Also, if a standard library module does the job just fine, then you're right, there's no need for other libraries to exist.

Nonetheless this sparked a lot of discussion on The World Wide Web where people were quick to tut-tut, saying "see I told you standard libraries were bad" in so many words. This is a popular opinion that's been floating around, at least in the few places I frequent that are admittedly Rust heavy. It seemed the article was heralding the end of the standard library for Python, and then ultimately, all standard libraries ever.

Then the axe came down with PEP 594, which chronicled the end of beloved, heavily used libraries such as

  • nis
  • imp
  • and sunau

Seriously they made it sound like they were going to get rid of json and re. Why was there all this fanfare around deprecating old and unused libraries that no one has used for 20 years? And why did they make it sound like the entire concept of a batteries included standard library was a bad and harmful thing? Python being batteries included is arguably why it achieved the popularity it has today. I'd kill for a standard ssl library in Rust - it would solve a massive problem with TLS support in the Rust community.

Oh well, I'll pour one out for my boy, sunau.