Hacker News Reader

Comment by jdxcode | Hacker News Reader

Engineers learn languages because of success stories by other companies—not by the merits of the syntax or one-off tricks like compiling itself in 5 seconds.

The fact that Firefox uses Rust for the browser engine or that Dropbox uses Rust for the sync engine is the kind of thing that gets me interested.

View on HN

Corporate sponsorship is the new prereq for a successful language/framework in today's world. It didn't used to be that way but it is now. Bring on our new googlemozillaapplefacebookmicrosoft language/framework overlords.

I dont think this is new. C and C++ came from Bell Labs, for example. Cobol from the USA government, FORTRAN from IBM. Java from sun, javascript from netscape.

In general, they all had a big institution. I presume these provided a a few ingredients, like

* A strong urge to solve a real-world problem

* A cradle, where smart people had time and resources to build a basic core

* A broadcast platform as others would look up to the institution, and notice what they were doing

From the top of my head, perl and python don't fit in this pattern, so its not universal, but I would dare to call it common.

That's a dangerously general statement, and it does not apply to me (an engineer) for example. I learned a lot of languages because of the language itself, e.g. their type system.

Perhaps. But general as it is this explains what OP was asking: Why Rust is more popular when D has the same functionality.

There's always people that are fascinated by new languages but I feel this is likely a minority. I think people want to learn languages they know will be useful.

It may be dangerous to blindly trust that a language is good based on a company's success with that language, but a large company deciding to heavily invest resources and take a risk on a language is also evidence that the language has the designs to succeed.

That's based on the company, tech...yes, insurance..probably no ;)

I think you have to be updated. Rust has had success in countless places after Firefox/Dropbox.

With that being said fast compilation times are amazing and must be a default in new languages. If not they should follow an approach like Dart where there is a VM for development and also a AOT compilation support for producing binaries.

There’s ongoing work for a fast debug back end called Cranelift. It emits code 20-50% faster than LLVM while being less optimised.

Explained further here - https://jason-williams.co.uk/a-possible-new-backend-for-rust

Yes, I have read about this. I follow Rust very closely. This is a good trade off while developing for speed.

D has a fast backend that omits assembly- it's DMD.

I didn't imply in the slightest those are the only success stories

Compiling extremely quickly will change your development experience to be like using a scripting language. As far as I know, D can still compile itself- and the entire standard library- in less than 5 seconds, with parallelism turned off.

I've never heard of a commercial success built off the back of Racket or J but happily went through a book on Racket and would love to do the same for J.

And yet almost nobody uses Racket and J. I think they were talking about more wide-spread use.

Uhh, that is quite wrong, lots of company's use racket, Haskell and even start new projects with Ocaml

Many schools around the world use Racket.

The financial industry makes heavy use of K, which is closely related to J. It's not open source or free to use for commercial use, however.

Then I recommend https://www.jsoftware.com/help/jforc/contents.htm

Thank you! I just bookmarked it.

Are you implying that D is slow at compiling regular code but fast at compiling itself, because of some cleverness?

Is that actually true?

Lol, the person you're replying to never said, nor implied that.

I'm not trying to knock you, but your comment reminds me of the type of exchanges where one person says says "men do a lot of work", and the other person replies "are you saying women are lazy?"

> one-off tricks like compiling itself in 5 seconds.

I took that as saying D could compile itself in 5 seconds too. I was impressed. Bubble bursted.

I'm pretty sure that was a knock at go

I'm not saying that at all. But looking at a feature matrix or hearing from a fan what the great parts of a language are isn't enough to convince me to adopt it.

If it's been around for 2 decades and hardly has any adoption by large companies[0] I personally am not willing to learn it (unless that changes).

I'm not necessarily using adoption as a proxy for quality either. If there isn't any meaningful adoption it's not like I can join a company to write it or have an easy time convincing my manager to try it. It could be the greatest language on earth but if I can't write it then it doesn't matter.

[0]: https://dlang.org/orgs-using-d.html

Reply on news.ycombinator.com