Hacker News Reader

Comment by RyanZAG | Hacker News Reader

The idea of Dart is great - but the 'Google' brand attached to it is very worrying. Dart feels like exactly the type of service that Google will just drop in the future if it doesn't get popular quickly enough, and with only Google supporting it, that makes any code written it in a huge risk. Google's track record is the single clear problem facing Dart at present. Personally, I've been holding off on Dart for exactly that reason.

I have no idea how Google could get around that - maybe a written promise to provide support for Dart for X years?

EDIT: Another similar worry: if Dart really does take off, will Google not just drop support for other browsers besides Chrome from stock Dart?

View on HN

As someone professionally responsible for maintaining and extending a legacy GWT project, I couldn't agree more with this view.

I'm no longer willing to tie my career to google's fickle attention span. Particularly since as I understand it the dart team has some of the same personnel from the old gwt team. They are masters at the art of soft deprecation through neglect.

Wait, what? My sympathies for maintaining legacy GWT, but that isn't at all equivalent to Dart except in that they both have JS compile targets. Go is at least a somewhat more reasonable comparison, but still isn't great. The biggest problem with any legacy GWT project is all the widgets people loved to make and use which grafted unnaturally to JS and the DOM, and certainly not the code transpiler, which is all that really matters here.

Dart is lead by Lars Bak out of Aarhus, which IIRC wasn't even a Google office location before they hired Lars to do V8. Even if some people transfered over from GWT, I doubt it would be many, and I don't see how it would have substantial impact with someone like Lars at the top.

And it's been 7 years since GWT was released. Yes, yes, enterprise needs whatever, but many web technologies fade away over time. It sucks if you were an early adopter and were heavily invested, but at a certain threshold, no one cares anymore. Is YUI's nearly negligible update rate a sign of Yahoo's fickle attention span? No, because no one uses it anymore (sorry YUI users), and at a certain point you have to cut your losses. If the community is self sustaining enough, it can pick up the slack with these open source frameworks, but if it's not, I really can't manage any indignation. It happens sometimes.

I'm not advocating using Dart in production (I would advocate just learning JS if you're targeting browsers anyway...), but this critique is specious at best, not least because the SDK is apparently just entering beta, so it says right on the tin not to found your production pipeline on it.

GWT may have been released 7 years ago, but it isn't like it was abandoned 5 years ago and the last stragglers have been dying off since then. The release tempo didn't fall off a cliff until the end of 2011, which just so happens to be when Dart was announced.

Not that there weren't signs of ADD before that, what with the dizzying changes in how-things-ought-to-be-done with each point release. But an attendee of 2011 Google IO could quite reasonably have come away with the impression that as far as Google was concerned the future of GWT was glorious. (see http://www.google.com/events/io/2011/sessions.html)

Even in this age of iterating, pivots, and agility, two years is a quick turnaround from breathless evangelizing to dead and buried.

Since GWT is open source, Google 'dropping' it (many largish Google projects use GWT and they still commit code) is not the death knell for GWT based projects. For the same reason it is being actively developed and has a steering committee and regular participation of googlers. It is also 'probably the best cross-compiled JS framework in the world', to borrow a phrase from the beer industry. Full-featured and works today, not 'maybe in five years' which is what Dart promises (I'll be the first to jump onto it when/if that time comes).

> I'm not advocating using Dart in production

then why anybody would want to use it? as much as i dislike MS , their TypeScript language makes more sense than Dart which feels DOA already.

Well I don't see why my advocacy would be the dealbreaker either way :) Unless you mean if it's not to be used in production, then what's the point? It's a new language; that's how they work. I wouldn't recommend people use Rust in production yet either, but that says nothing about its future.

I agree with you on Typescript, though I don't see it as any more interesting than Closure Compiler annotations plus one of the various ES6 transpilers. It's nice to have those in one package, but I don't really use or want the transpilers normally, so I feel no draw to them now.

As for Dart, I've never totally understood its intended place and how it was expected to succeed there (though we'll see, I guess), but as for its use in practice, philosophically I now just think of it as another Coffeescript, and so I'm at peace with the idea that people want to use it.

Don't forget that Dart is not just a language, it's also a tool ecosystem, with a virtual machine that is already faster than V8 in some key benchmarks.

This session from I/O by Kasper Lund and Lars Bak is well worth a watch: https://developers.google.com/events/io/sessions/324431687

In it they explain that they've pretty much reached the peak of where they can get to in terms of performance with V8 (no more big jumps in performance), but by designing a language that is easier to optimize for, they've already made the Dart VM faster.

How can you compare a language that introduces a bunch of syntactic sugar over JavaScript semantics (CoffeeScript) with a language that brings a cleaned and streamlined semantics (Dart)?

Only in the sense that I could have also compared it to any of the other languages that compile to Javascript. It was not meant as a commentary on the value of the language design. In fact, it was the opposite: Dart's mere existence doesn't mean I have to care about if people should be writing pure Javascript or worry about any potential Javascript existential crises Dart might raise. And that's great.

It was a process to stop worrying and love the bomb -- so many HN articles proselytizing and condemning Coffeescript (and others) -- but I believe the healing started with the sheer popularity of Coffeescript, and then normalized when Brendan Eich stood on stage with Jeremy Ashkenas and told everyone that compiling to Javascript was A-OK in his book.

So thinking of these languages in those terms is the reason I don't really care about the horse race and I'm ok with people writing in Dart if they want to. My domain will be fine, and might be improved as they find interesting things (syntax, developer tools, whatever) exploring a slightly different path toward a web scripting language.

(though I'll continue to be annoyed if some article on web technology puts all its examples in X where X is not Javascript :)

I don't think this is a problem. It's rather uncommon for languages to be supported in the first place. Rather, independent contractors become experts in the language and pick up the slack.

> Another similar worry: if Dart really does take off, will Google not just drop support for other browsers besides Chrome from stock Dart?

If Dart, an open-source toolchain, takes off (e.g., becomes wildly popular with developers) and then Google chooses to limit future Dart versions to only work with Chrome, I would expect a fork to become popular in a heartbeat. I mean, look at other popular open source projects whose sponsor/maintainer has done things unwelcome by the community (e.g., OpenOffice, MySQL.)

"Dart feels like exactly the type of service that Google will just drop in the future"

The difference is that Dart is not a service, and cannot just be turned off. It's an open-source language, and "at worst" could be community maintained.

This is exactly why I'm not touching it. At least with GO they are using it in all there servers, so you know they will be supporting it for a while.

Two corrections: 1) We are not using Go in all of our servers. Java, C++, and Python are very prominent. Go is just getting started internally. 2) Dart is being used internally, we won't be dropping it.

Can you say which internal projects are using Dart at Google? Were they new systems, rewrites, or evolutions of existing JS systems?

They got some patches from the finance department and Google were really insistent on adding big integers to JS so it looks like they're using it for currency. Woo... exciting.

Dart is an unilateral project aiming to: "ultimate goal is to replace JavaScript", "rescue browsers from JavaScript" or "Google preps Dart to challenge JavaScript". Being open source here, is not good enough.

Reply on news.ycombinator.com