Hacker News Reader

Comment by Gigachad | Hacker News Reader

[flagged]

RoR peaked in 2013.

I think you may be referring to Matz, creator of Ruby, who has made a stance against strong types in Ruby. We do have Sorbet, but that doesn’t go far enough for strong type enthusiasts. DHH is only a secondary voice in this area.

Nah, when DHH was asked on Twitter what he thinks about having rails typed he responded “Absolutely not” and never gave any more reasoning.

I don't see how DHH has any say in the matter. It's the language which is statically or dynamically typed, not the framework.

Typing in Ruby will be an optional thing. You can already do it now. It’s much like opting in to typescript. If rails won’t add the types, it’s going to hold back typed ruby.

Maybe he didn’t get the memo that we’ve tried dynamic types and it’s not that great of an idea outside of some edge cases

Yes that’s why everybody collectively moved on to static types exclusively 30 years ago and we all lived happily ever after because there was never a point where dynamic typing made more sense or was anything but inferior. The end.

My theory is that types were the baby thrown out with the bathwater. Dynamic languages changed a whole lot more than just types. The older typed languages were a huge pain to work with, dynamic languages fixed a lot of that and added standard libraries that do a huge amount of work for you.

Now TypeScript and Rust are showing typed languages can be just as productive, if not more productive than Ruby/Python/JS.

I don’t buy it. Static languages have certainly become more ergonomic, but if the evidence was really so overwhelmingly clear, everybody would have switched by now. To wit, regarding your comment about Rust, I just read this article earlier today: https://mdwdotla.medium.com/using-rust-at-a-startup-a-cautio...

> Static languages have certainly become more ergonomic, but if the evidence was really so overwhelmingly clear, everybody would have switched by now.

Change happens slowly in our industry, but Ruby very much feels in decline. More telling IMO is that every major new language to emerge this millennium has been typed.

Clojure and Elixir are dynamically typed as are the two most popular languages - JS and Python.

TypeScript seems to have by-and-large "won" in the JavaScript ecosystem; I can think of exactly one senior+ developer I've run into in the last few years who's expressed a preference for JS over TS and, similarly, I can think of only one greenfield purely-JavaScript project I've seen started in the last three years; that project was built in that way largely out of ideological objection to having a build pipeline. (I think it was a mistake.) It isn't a majority yet, but it seems inevitable that the future of JavaScript among non-laggard populations is TypeScript or whatever will supplant it--and whatever will supplant it will not discard types.

As for Python, adoption is slower, but for me at least, Python typings are becoming the reason to consider using Python in 2022 and I won't use a library without them. And the worst part of Elixir, from where I stand, is that typespecs are bitterly underused. I have some hopes for something like Gleam to make a backwards impact upon Elixir, because my experience with development teams trying to bang on Elixir services without it is pretty grim.

Typescript won huh? You're using your own anecdotal experience to support your statement. I'm looking at random Javascript questions on Stackoverflow now - there's a gaziilion of them from the last hour. None of them is Typescript. So plenty of people still write JS.

https://stackoverflow.com/questions/tagged/javascript

I use Typescript and like it btw but I just don't agree with you here...

I'm sure there are plenty of projects created in JavaScript. I'm not sure there are plenty of projects created in JavaScript in what I think we can all generally agree is leading-edge to mainline tech.

There will always be laggards and there will always be novices, but it is so incredibly rare to see a library without typings or a senior developer not start with TypeScript almost without thinking that I think it is at worst only a matter of time.

As an aside, why are all of your posts written so combatively? I looked at your most recent page of comments and if I'd not already responded to you, I wouldn't have. I certainly don't shy away from being pointy, but not about things as ultimately meaningless as programming languages.

I'll stay on topic. If I was somewhat combative towards you it is because you made very big statement which in my opinion are simply inaccurate. My tone was somewhat harsh but your comments are in my view simply wrong. They also give the air of a somewhat condescending know it all (of course Typescript won!) which is a bit childish and can be very annoying to read, and this is going on quite a lot here with many other people who've seen the tech light and now need to hammer it into us commoners heads. Anyway looks like we've beat this horse to the death.

Just about everyone has switched from JS to TS.

When given the choice of the exact same language with and without types, it seems pretty clear that typed languages win.

>>> Just about everyone has switched from JS to TS

You are mistaken. While TS is popular it's nowhere near as popular as JS. In fact, just looking at online community counts, I'd say it's 1/2 as popular. Will it continue to gain? I'll never bet against JS.

I don’t think that’s correct and you’re reading too much into popular blog posts (“we tried typescript and decided it wasn’t for us” doesn’t make for a good article). I don’t have any hard data either, but since we’re swapping anecdotes it’s my impression that typed pythons uptake has been very meh outside of multi-million line codebases (and is basically not seen in ML work!), and typed racket and typed clojure are basically unused in practice

Problem is switching programming languages is a big deal. It is a big bet to take. Static typing is probably not enough of an advantage alone but has to be taken into consideration with all the other tradeoffs.

By what measure of "productive"? If you mean developer productivity Rust is, by the admission of its own community, the slowest development environment. Rust is great where memory safety is paramount but it doesn't compete in the same playing field as Rails for CRUD apps.

I would say that with Rust you can be "more" productive by deploying apps that requires less maintenance as for example Ruby or JS app, also a pain with JS/Ruby is updating dependencies, that's a very scary thing to do if you are far behind, while with Rust you can update and instantly see where your code broke.

But I do agree that with Rails you can have quickly a more functional app, but that will be sooner or later not true anymore.

Can C# enter the chat? “I have been pretty good all along”. F# too!

I admit C# suffers from the OO fetish roots, but it has pretty nice language features now so I almost never need to use inheritance.

(I might be very wrong as I don't have time to check every statement):

Actually the theory of static typing appeared _after_ the theory of dynamic typing.

So the history seems to be:

1. dynamic types

2. static types

Thus you can say some people didn' get the memo that we tried static typing and it not a panacea. It fits very well some cases and it does not fit some other cases.

Dynamic typing is here to stay, so is static typing.

And when referring to Ruby here is how to think about it:

Ruby is strongly typed, but it does not use static type checking thus it is defined as dynamically typed.

Reply on news.ycombinator.com