3.9 KiB
+++ title = "Python Losing the 2nd Place on Loved Languages: A Thought" date = 2020-06-09
[taxonomies] tags = ["python", "typescript", "stackoverflow", "survey"] +++
On this year StackOverflow Survey, Python lost its 2nd place in the "Most Loved Languages" ranking to TypeScript. On our Python group, people started wondering why. And I just thought it would be nice to post my thoughts on the situation.
Before anything, I may have do make a disclaimer saying that I do love Python, and I think it is a great language 'cause it is very concise but also very expressive. I don't think any other language get close to Python in doing those points so well.
Disclaimer done, let me explain why I think Python lost its second place in the hearts of developers.
Types
One of the first thoughts one can think when we talk about "losing a position" in any rank is "because it got worse". But I really don't think this is what happened here, and what actually happened is that TypeScript shown developers something better.
TypeScript is aimed to JavaScript development, a zone where anything goes, types are very flexible and magical -- to the point what most of things actually produce a Wat. By using types, making sure you're not adding a string to an array, a lot of problems suddenly disappear. It is no silver bullet, but it prevents a whole class of issues that would popup in production.
So, in an environment that chaotic things happen, suddenly you have something that puts order back and you gotta love it for that. Surely, it's not just types that make TypeScript more loved than JavaScript, but it is partially why.
(Just to add to this point: Rust is still the most loved language 5 years in a row, and it uses some pretty strong typing, close do Haskell, with added syntax closer to what most languages use.)
I still believe that, in the long run, type hinting can fill the gap for Python to reach the proper position. I'd love to use it to add hinting in all functions and then have something that I could set the level of checking on different environments: make the application crash if the function was called with the wrong types on development environment; just log (logging or stderr) on staging; and do absolutely nothing on production. That would allow me to use the best of both worlds: Dynamic when developing, but static when testing.
Python 2 is dead
One of the things that happened to Python in the beginning of this year, when the survey was done, was that Python 2 became unsupported. Surely, that doesn't mean any Python 2 installation would stop working, but it gave an extra jump to porting things to Python 3.
And, even with all the backporting and improvements in Python 3, it was not a smooth sail. That change, that forced change, may have let some Python devs with some bitter taste about the language. No one wants to just fix changes in the language, when the way things work must still work, quirks and all.
Hype
Python is not a hyped language anymore.
Sure, it is still a reference for machine learning and related fields, but the once thriving environment of web dev was taking by the hype of other languages.
Python is not the hot stuff on web dev anymore. And because it is not the hot stuff, people don't want the old stuff; the old stuff is not cool anymore, so they don't like it anymore.
Related: Just because something has hype, it doesn't mean it is better; it just makes the non-hyped stuff "not better", even if there was no change at all in the latest.
Side-point
One point not raised by anyone in the group: Although Python lost its second place in the "Most Loved" ranking, it is still the most wanted language -- meaning, it is the language most developers want to learn. If the general feeling was "Python sucks!", I pretty much doubt the want would still be representative -- and Python have a large lead compared to JavaScript.