Hacker News Reader

Comment by cryptos | Hacker News Reader

Why Python at all? About 10 years ago I liked Python a lot (and still like it in principle) and felt very productive compared to, say, Java. Java was full of inconvenience, XML, bloated frameworks and all that. But today you can use Kotlin, that is in my opinion even nicer than Python, with performant frameworks (e. g. Quarkus or Ktor) on the super fast JVM.

I don't want to start a language war, but maybe Python is not the first choice for their requirements.

View on HN

I can code comfortably in Python, Java, JavaScript and to some extent C/C++. In the last 4 years I have been using mostly Python for various reasons (Machine Learning, OS automation, web scraping ...). Compared to the other languages, Python feels lighter and faster to write to the extent that it rarely interrupts my flow of thoughts. Now and then I have to code in JavaScript (frontend), C/C++ (embedded / low level optimizations) and Java (maintenance) and my feelings are:

- JavaScript still feels messy

- C/C++ is complex, but it is often offset by the complexity/needs of the project (e.g. in embedded)

- Java has become kind of bloated with all the new stuff

So as others have already mentioned, development productivity is in many cases far more important than speed of code. In my ~20y long career I have rarely seen a project failure due to runtime performance. Most of them failed due to speed/agility of development iterations and also project/product management issues (bad fit, unrealistic project plan, lack of focus and customer feedback).

That said, if I need to look for another language due to performance, that would probably be Rust.

I also notice I spend less time thinking "how do I do X in this language" with Python and just do it. Maybe not the best implementation of X, but that doesn't really matter because I implemented X, Y and Z while someone else got bogged down in their "better" language and barely got X across the finish line.

I agree with your general take on developer productivity, but I don't feel that modern JS is significantly messier than Python, at least not to a level where it significantly impacts productivity (I'd rather avoid a debate on the abyssal depths of the language, eg, type coercion)

I feel about the same amount of grievances with both. For instance I dislike Python's async and functional semantics (list(map(lambda n...). But it has a much better standard library overall, a lot of the scripting syntax semantics (file opening, requests, etc.) feel cleaner, etc.

I'm more versed in JS but as the knowledge curves converge in months or years to come I don't think I'll be significantly more productive in Python.

I know you were expressing an entirely subjective opinion, probably contingent on the amount of day-to-day practice you have with each language (you say you code JS in the front-end now and then... I write a lot of it), but I still wanted to offer my counterargument.

> list(map(lambda n...

List comprehensions are much better for this. Functional doesn't mean you have to use a function call. If you can use the paradigm with literal syntax, just do so.

I never said functional means you necessarily have to use a function call, and I also understand list comprehensions are semantically functional, but sometimes you do need to use a function call.

They don't really compose nicely. (At least from my point of view, as I prefer fluent interfaces, eg. those you usually find in Rust/Scala/Java.)

Thanks for the JS perspctive. Since my frontend work mostly involves hacking together prototypes and trying to integrate some JS libs in as quickly as possible, I guess I am kind of biased.

JS feels very mess to me too. I think the main reason is it’s so variable - things like modules - whereas Python everything is mostly just so consistently... pythonic.

I agree - if you have any chance at the slightest volume (and based on this we're talking 100 concurrent users, not a million) then you are handicapping yourself if you're stuck on a single thread.

Personally I would suggest Elixir (and the Phoenix web framework). It's fully 'parallel by default', will happily saturate all your cores and serve tens of thousands at the same time. The language is simple, the framework is good, although obviously not as many people know Elixir as Python.

I feel like when you need to create value and make a product, better to do it in a language you know. Product value is not directly proportional to product performance or raw HTTP request serving time.

I for instance know a bit of F#/C# and Java, but would probably pick Python to make a new product just to remove that mental barrier of not having my lack of language knowledge in the way of things

Reply on news.ycombinator.com