21–30 of 369 posts
I'm very excited about Elixir for example. There is a new Elixir blog post/article/github repo posted on HN every day, you would think that Elixir is the new cool thing. Well, not easy to find companies that actually use it.
Reading HN, you'd think Java was dying (dead?) too. But it's not. It's an established language with a lot of very senior people still using it. It has it's issues for sure, but it also has a ton of libraries and support. Python I'd say is in a similar boat. Strong and established with tons of libraries, not popular amongst the kids. Edit: Apparently my use of "kids" as a sarcastic way of referring to those who won't…
Java has gotten significantly better in recent years across the board. Python it's a mixed bag due to the 2 to 3 incompatibility transition, one step forward, one step back and some step are changes but not improvements.
I think it's HN echo chamber. A lot of articles here about new exciting technologies, but when you look at job postings for example this is not the same song. I'm very excited about Elixir for example. There is a new Elixir blog post/article/github repo posted on HN every day, you would think that Elixir is the new cool thing. Well, not easy to find companies that actually use it.
Python occupies a niche that isn't going away any time soon: making it easy and natural to write readable, straightforward, more-or-less imperative, slightly boring code of the type you learned in CS 101. This is still a very practical way to solve many problems and I'd wager for most programmers it's still the easiest way to do things. Maybe it will always be. It's hard to imagine there'll be a generation of program…
Why can't Julia fill that role?
Python was doing just fine until Python's little tin god came up with Python 3 and tried to drive a stake through the heart of Python 2. That backfired, badly; six years on, Python 2 still has far more production use than Python 3. (No, it's not about Unicode. Python 2.6 and later do Unicode just fine. I've written sizeable all-Unicode systems in Python 2.6. It's just that in 3.x, it's the default.)
Coming up next, Python 4, with typing syntax that isn't checked. Ugly typing syntax - some in comments, some as annotations. There's an argument for optional type declarations that are checked, but unchecked declarations are trouble waiting to happen.
The PyPy crowd has brought off a miracle - an optimizing compiler for Python that supports almost all the dynamism of CPython. It took a long time, but they did it. CPython should be retired now, or made to conform to PyPy. But no; the little tin god insists that his CPython defines the language. The PyPy developers struggle to keep up.
The language isn't the worst problem. It's the wildly variable quality of the libraries. The great thing about Go is that it has libraries which are used internally by Google, and thus have been executed billions of times. Python's PyPi (formerly Cheese Shop) has no quality control, and many versions of very similar libraries, each with different bugs. There's no convergence.
Python occupies a niche that isn't going away any time soon: making it easy and natural to write readable, straightforward, more-or-less imperative, slightly boring code of the type you learned in CS 101. This is still a very practical way to solve many problems and I'd wager for most programmers it's still the easiest way to do things. Maybe it will always be. It's hard to imagine there'll be a generation of program…
Why can't Julia fill that role?
I play around with NodeJS and front end frameworks like React for my own side projects, but all of my paying client work is made with good ol Django. It just does so many things well and is very mature. I recommend for production code using libraries that have a strong community and have been around for a while so that they don't keep changing all the time.
You probably still want Django to handle server-side rendering for most of your site (unless you're a masochist). Django Rest Framework for the API, and React (or whatever) to handle the app-like parts.