The advancements in compilation and JIT-compilation have made interpreted languages pointless.
Except for inertia-related reasons such as existing ecosystem, why would you use Python over Julia? It has all of its advantages and more and compiles to efficient binaries instead of redoing runtime maintenance work every time.
1) I like class objects and the ownership of their methods. Julia doesn’t offer this.
2) JIT doesn’t really gain / offer me that for the work I do (yes it is faster but nowhere near make-or-break)
3) In my field, python is already rare enough (SAS, Stata, R are the standard). Most have no idea what Julia is (so will be hard to get collaborators to buy-in).
4) Related to 3, most in my field favor familiarity with a software over speed.
5) If speed was what I really cared about, why not just go to C, C++, Rust, etc. and skip over Julia entirely?
6) Really can’t discount the existing ecosystem. Julia has a long way to go
@PausalZ oh yeah, I don’t discount the existing ecosystem nor claim that you should switch to Julia immediately. Yes, inertia-related reasons like the existing community, knowledge base, libraries and plain familiarity are very important and keep languages like Python alive even when worthy successors are present. This is not a “switch to Julia” post, it’s a “interpreted languages have become an artifact of the simpler past” post.
Nevertheless, I’m happy to inform you that Julia’s type and method overloading system provides all of the necessities for OOP too, and it’s not slower than C, C++ of Rust; with only the latter having the advantage of nearly runtime-free memory safety, at the cost of many restrictions. A properly done Julia program is as efficient as a properly done C program that also uses adequate high-speed calculation libraries for the platform, at a fraction of the required effort.
