Did Google chicken out?

[Mayuresh Kathe]

Hello,

Not belittling anyone, but, upon reading a lot about Dart's motivation and older target, it feels like Google killed off the Dart VM in the browser project because they chickened out while "playing chicken" against the rest of the industry. Blah, what a shame!

I think I am just too intrigued and impressed with Dart to want to work with it long term, if not the browser, then the server it will be for me.

Just hoping that the Dart VM on the server continues to be innovated upon and developed as long as Dart exists.

Best,

~Mayuresh

PS: again, an apology if I have veered off-topic.

[ravi teja]

Dart VM not being in Chrome doesn't mean end of the world for Dart on browser. Actually it is a good thing. To me, Dart is more exciting on the browser side than the server side. It makes web app development really really easier.

My own experience: I am able to complete projects using dart and polymer 1/4th of the time i took earlier with Javascript. Now, suddenly web development is more fun than ever thanks to Dart.

I wont call it chickening out. I would call it a great strategic decision. There is no way you can convince browsers vendors like Mozilla, Microsoft and Apple to add new VM. They have their own agendas (if you ask me, quite stupid ones :P).

--
For other discussions, see https://groups.google.com/a/dartlang.org/
 
For HOWTO questions, visit http://stackoverflow.com/tags/dart
 
To file a bug report or feature request, go to http://www.dartbug.com/new

To unsubscribe from this group and stop receiving emails from it, send an email to misc+uns...@dartlang.org.

[Mayuresh Kathe]

Looks like JavaScript has indeed become the byte-code of the Internet, with languages like Dart and TypeScript becoming favorable for writing full-blown applications due to benefits like decent structuring and excellent support for tooling.

I used to develop for the Web around 1997, and then backed away from it due to the disaster that is JavaScript. Hoped something like Dart would have come along to save the day, but well, things really don't change all that much, don't they?

I just took the conscious decision to stay away from Web front-end development, and run with Dart for the server-end stuff, hope things work out for me.

Yeah, and I agree, Mozilla and the others you mentioned do have pretty strange agendas.  :-)

~Mayuresh

[Bob Nystrom]

On Wed, May 6, 2015 at 8:50 AM, ravi teja <teja...@gmail.com> wrote:

My own experience: I am able to complete projects using dart and polymer 1/4th of the time i took earlier with Javascript. Now, suddenly web development is more fun than ever thanks to Dart.

This, to me, has always been the compelling motivation for Dart. A VM is nice too, but improved programmer productivity is what gets me up in the morning.

Cheers!

- bob

[Kirth Gersen]

Too soon to tell if it's a "great strategic decision". What is sure is that's isn't not a bald/'moonshot' decision, it's not "Googlish" and that is why a lot of people, including me, are upset about this decision.

Fact is Dart audience is very low. I watched all of the latest Dart Submit live online. I was surprised to see the very small room, half of it filled with Googlers and with barely 250ish live online viewers (worldwide)...then, except for the Sky session which was largely 'hyped' in the press, all the Dart Submit sessions on Youtube are, atm, mostly all under 2000 views worldwide.

That is way more worrying than them dropping the VM for the web and could explain the recent strategic changes.

As for the dartVM into Chrome, I think adding it, dropping V8 and providing a JS2Dart JIT compiler into the Chrome would have been the 'bold'/moonshot move. But V8 is also from Google...and much more used, stable and optimized.

[Mayuresh Kathe]

Kirth, the numbers you've mentioned make the whole picture very bleak and scary for Dart.

Thanks for the tip, I now need to rethink my decision to run with Dart in the long term.

What if suddenly it gets killed off completely? Won't my investment in time and resource be a total waste?

[Kirth Gersen]

What is sure is that we have no guarantee.

We're all there atm: rethinking our decision to use Dart or not.

The VM in Chrome would have reassured us because Chrome is strategic for Google so having Dart strongly tied to Chrome would have been a clear, long term commitment.

All we have now is their word , and only from the Dart Team and some internal users like Ads, not from higher management at Google (Sundar or above for instance).

And they're working on providing an 'exit strategy' tool: the dev_compiler ... ;) (I'm mildly joking here... but it could be used to move away from Dart...).

Cloudy and uncertain the future is.

On Wednesday, May 6, 2015 at 6:21:55 PM UTC+2, Mayuresh Kathe wrote:

Kirth, the numbers you've mentioned make the whole picture very bleak and scary for Dart.

Thanks for the tip, I now need to rethink my decision to run with Dart in the long term.

What if suddenly it gets killed off completely? Won't my investment in time and resource be a total waste?

[Frank Pepermans]

It's so hard right now to compete with JavaScript. It's as if Dart were standing all the way in the back of a <insert fav artist here> concert, crying out for people to look the other way.

The mobile story is good, maybe it can get a better foothold there and retry the browser story again after more people start looking at it.

If Dart will soon be able to compile to js *as if natively written js* then why not try out Dart at the company, worst case you could always continue working on the JS code?

--

[Mayuresh Kathe]

Hmnn, in that case, it would make better sense to focus on TypeScript.

Microsoft has never made any grand play to turn it into a "replacement" to JavaScript and then backtracked.

--

[ravi teja]

>> As for the dartVM into Chrome, I think adding it, dropping V8 and providing a JS2Dart JIT compiler into the Chrome would have been the 'bold'/moonshot move. But V8 is also from Google...and much more used, stable and optimized.

That would be the stupidest idea ever IMO :P. That would slow down existing websites. Sounds like the greatest way to kill chrome :P.

I don't know why people take Dart being in Chrome so serious. If Dart is included in chrome, so what? I don't see a way it will be included in Firefox, Safari, Internet Explorer.

Dart is cleaner and easier to work with. It wins huge in maintenance.

If you want some moonshots, how about:

1) Observatory

2) Sky

3) Fletch

I just can't wait for these three projects to attain completion. It would make cross-platform mobile development totally fun.

--

[Joel Trottier-Hébert]

@Kirth I don't see how the dev_compiler is an 'exit strategy', it looks to me that Dart is going wide public with this.

They said it well 

Export your Dart library in JavaScript (SYODLAJ: Ship your own Dart Library as JavaScript)

@Frank Dart has always been the "unloved person"..

I remember as far as 2011 in Dart Preview http://googlecode.blogspot.ca/2011/10/dart-language-for-structured-web.html

People didn't like the idea, why would they like it more today?

Don't get me wrong, but I think Dart haters are [most of the time] some old school people coding in [something, often Javascript] that fear for their future. I'm not saying they're wrong or right, just my general impression on Dart's image in the wild.

Joel Trottier-Hebert

[Lex Berezhny]

Google did do a moonshot: Sky.

If you believe for a minute that mobile devices are going to be more and more prevalent going forward (fact) and there is a lot of research and analytics showing that most mobile users prefer mobile app vs visiting equivalent website (fact). This basically means that iOS/Android are the primary content delivery platforms for many people today and more tomorrow. The web browser as we know it is becoming less pervasive.

We as developers or tech workers spend a lot of time in the browser but we are not the entire population of the world. Most people on this planet do not use a web browser (in fact they don't even have internet access)... even those that do have desktops/laptops at home might spend more time in their mobile apps (Facebook, etc) than using browsers.

There is more untapped potential (in terms of humans on this planet you can reach) in making mobile apps than making websites.

So, Google could have spent a lot of effort adding Dart VM to Chrome...

Or they could add Dart VM to Android (Sky) along with a newer/faster rendering pipeline to target the rest of untapped humanity on this planet.

If you remember, the arguments made at Dart Summit and a few other places for why Dart is the way it is and why it's better than ES6 is that Dart doesn't carry the baggage of JavaScript.

I think Sky follows in that same spirit. If Dart is free from all previous baggage, then it only makes sense that the main application framework for Dart should be free from previous baggage, hence Sky.

Web browser as platform is not that exciting anymore. There are more web tools and frameworks than you can learn in a life time. I think we're finally getting to the point where we need to break free and start fresh. Web browsers are best to use on a large screen with mouse and keyboard and clearly this isn't the future...

If you like the idea of innovation on the browser then use ES6, I think that's the best that can be done on such an old platform. Besides, deep down we all knew Dart VM wasn't going to be added to Firefox, Safari and IE.

I think Google decided to abandon adding Dart to browsers for the same reason they abandoned their medical projects. It's not efficient to fight such an uphill battle against entrenched interests and bureaucracy. And if there is no chance of Dart getting into Firefox, Safari and IE, what's the point?

 - lex

[Joel Trottier-Hébert]

... also, Javascript 5 years ago isn't the same as Javascript today, or tomorrow. ES6 shows a lot of great improvements, and therefore transpiling to ES6 is something desirable now I think.

Joel Trottier-Hebert

--

[Mayuresh Kathe]

Lex, I had not idea about "Project Sky", it does look VERY impressive.

Thanks for sharing those details.

--

[Jim Trainor]

Dart as productive, modern, alternative to JS for web development = big win.

Fletch as execution environment for client side dart code on iOS and Android = big win

Sky on the horizon (pun intended) = big win

No Dart VM in Chrome = small loss (or no loss at all for those of us that would have deployed to JS even on a Dart VM enabled Chrome).

[Joao Pedrosa]

Hi,

On the server Dart has to compete with Java, Go, Node.JS, and even with Rails... :-)

That said, Dart has a neat project in Fletch that I hope will catch on and help it on the server. Fletch supports concurrency well and it could give us a different kind of building block. But they talk about it as an experiment that could eventually be incorporated into the default VM. I'm not so sure about how they would do that... But...

I enjoyed watching some of the talks of the Dart conference. They showed that the Dart developers have some of the same concerns as us.

I hope that our current anxiety will one day pay off. My feeling is that we should not take Dart for granted. For instance, I gave a quick try to the IntelliJ Community Edition and also quickly removed it. My hope is that Dart will one day have its editor coded in itself. Maybe that's what some of them have in mind as well. Once Dart has that, the Sky would indeed be the limit.

Cheers,

Joao

[Frank Pepermans]

+1 after using Dart day to day for over 1.5 years, it bites to go back to the classic stacks.

The story is very nice and hopefully the job market will catch on soon enough

[Bob Nystrom]

On Wed, May 6, 2015 at 9:43 AM, Kirth Gersen <kirth...@gmail.com> wrote:

All we have now is their word , and only from the Dart Team and some internal users like Ads, not from higher management at Google (Sundar or above for instance).

The announcement quotes Scott Silver, the VP of Engineering for Ads. He is higher management. I can't go into details of our org chart, but Scott is quite high up it.

But, ultimately, I understand your concern. A programming language is one of the most deeply entrenched technology bets you have to make as a software person. Your entire product will use it as its media, and all of your employees will spend a ton of time loading it into their precious brain cells.

You want some confidence that you are making a safe choice. At the same time, you also want to take an opportunity to make a smart choice when a better language can give you an advantage over your competitors.

Programming languages are also hard here because the market demands that they be free. Programmers won't learn languages that you have to pay for. So you want to be absolutely sure you can depend on a product, but you also aren't financially supporting the people that make it.

I think this is a big part of why language evolution is so haphazard and slow: the market doesn't handle programming languages well. Basically, you just hope a big organization (Google: Go and Dart, Apple: Swift, Microsoft: C# and TypeScript, Heroku: Ruby, Dropbox: Python, etc.) will support a language without any immediate financial incentive to do so, and then rely on your trust of that organization.

Here are some signals that Dart is reliable:

• We have a lot of internal customers using it. Ads is the most important arm of Google because their products make most of the company's money. If Ads is happy using Dart, they are more than willing and able to keep funding it.
  
  
• The quantity of internal code written in Dart is growing quickly. Obviously, I can't get into numbers here, but it looks like internal adoption is going great.
  
  
• The Dart team is large and stable. We are pretty large for a Google team and over our history, I haven't seen any push to shrink it.
  

Cheers! 

- bob

[Kirth Gersen]

ty Bob.

Can we expect more positive 'signals' at I/O 2015 ?

I'm also worried there (yeah I worry a lot :)) because I see nowhere the word 'Dart' in the I/O 2015 schedule unlike in 2014 which had plenty Dart sessions.

[David Notik]

Thank you for emphatically, clearly and directly addressing our concerns. Though I'm sure it feels like you're repeating the same mantra over and over, it's needed for those of us out here making big, precarious bets. Some of us are investing ourselves in Dart without pay or assurances of any kind. Some are convincing our boss or team to adopt it. Your insights and statements of support go a long way.

Nobody likes to feel they're with the losing team. It can seem like the party is elsewhere (React), and ours is a little campfire in the woods. But I say judge Dart based on how effective you think it helps you be. Do you feel it helps you build and deploy the kinds of things you want to, more so than if you went a different direction?

None of us can compare every platform to quantifiably determine which we'll forever be most productive in, and that's why we look at all we can and then we make bets. Once we make that bet, the cost of changing that bet only increases, and that's the understandable source of our concerns. It's a healthy discussion, and I'm glad the team repeatedly engages.

Luckily, while our bet is a constant question, what's even more constant is the productivity we can all feel with a really great language and toolset, and the pace of change and innovation – on the community side with the summit, the rapidly advancing language and libraries, the improving debug and deploy story, the exciting push to unleash our Dart productivity on native apps, and more.

(Personally, I'd like to see a better answer with respect to a client side approach to web apps. Angular and Polymer are both for building SPAs, but React's fundamentally different approach has big benefits: its components can be rendered server or client side (big for initial load and SEO) and its Virtual DOM is highly performant (which really makes a difference on mobile). The latter is mitigated somewhat with a great native app story (though we still want performant apps in mobile browsers), but the former is a really big deal for many. I'd like to see the Dart team better speak to those.)

The story today is great, the story unfolding is even greater. Many of us are productive with Dart thanks to the team's excellent work. Let's build great products, and more enthusiasts will follow.

Whether or not the Dart VM is in Chrome does not change a thing, IMO.

[George Moschovitis]

Though I'm sure it feels like you're repeating the same mantra over and over, it's needed for those of us out here making big, precarious bets. 

+1 

[Bernardo]

On Thursday, 7 May 2015 02:44:14 UTC-3, David Notik wrote:

(Personally, I'd like to see a better answer with respect to a client side approach to web apps. Angular and Polymer are both for building SPAs, but React's fundamentally different approach has big benefits: its components can be rendered server or client side (big for initial load and SEO) and its Virtual DOM is highly performant (which really makes a difference on mobile). The latter is mitigated somewhat with a great native app story (though we still want performant apps in mobile browsers), but the former is a really big deal for many. I'd like to see the Dart team better speak to those.)

In the Dart Developer Summit there was a talk by Workiva. They ported a large javascript app using react to dart. They are using the react bindings that are on github[1]. So I think it's perfectly viable to use react now. 

Also the sky framework[2] is based on similar concepts as react.

[1] https://github.com/cleandart/react-dart

[2] https://github.com/domokit/sky_sdk/tree/master/packages/sky/lib/framework

[Bob Nystrom]

On Thu, May 7, 2015 at 6:36 AM, Bernardo <aux...@gmail.com> wrote:

In the Dart Developer Summit there was a talk by Workiva. They ported a large javascript app using react to dart. They are using the react bindings that are on github[1]. So I think it's perfectly viable to use react now. 

Also the sky framework[2] is based on similar concepts as react.

[1] https://github.com/cleandart/react-dart

[2] https://github.com/domokit/sky_sdk/tree/master/packages/sky/lib/framework

Exactly right. I look at React as a framework style, not a language choice. If the web world moves to React, the Dart ecosystem can as well.

Cheers!

- bob

[David Notik]

Absolutely, but Dart (as a community and set of activities, not just a language) offers up Angular and Polymer, and they're certainly the framework styles Google is behind. So that's what Dart communicates to web developers surveying the landscape for their next project. "If the web world moves to React" sounds like we'll be last to that party, when we probably should be asking the hard questions and innovating all along. The official activity around a React-inspired approach to native apps is great. I'd like to see word from the team about framework styles, current approach versus the direction others seem to be going, as well.

--
Woven
http://woven.co
206-351-3948

@DaveNotik

--

[Zied Hamdi]

It is simply a political decision to interrupt chrome dart native support : the web community had a paranoiac reaction saying that dart wants to kill js, and Google wants to rule the world (this kind of fear propaganda). Si since the objective of dart is the ojective of chrome: to make the web faster and a better experience (both for developers and 'mainly for' users). Interrupting chrome native support was not such an issue, after all, the core of the dart language is its javascript multibrowser compiler, the native support is just a performance "hack" that can be removed without breaking the core value

I'm pretty sure that when dart will be widely used, it will be a community request to have it supported natively on android, browsers and why not iOs, blackberry etc... after all, it's just a language!!!

Cordialement,

Zied Hamdi, Gérant  

http://1vu.fr - +216 5206 8582

Many vocations, one view 

--

[Don Olmstead]

When talking with the sky guys at the summit they said that sky speaks web components as well it just wasn't in the demo they we're showing.

https://github.com/domokit/mojo/tree/master/sky/tests/custom-elements

[Grant Jason]

@Joel 

Clearly you love Dart and so do I. 

But at least be man (clear-headed) enough to admit that Google has decided to reduce it's investment in Dart quite drastically. 

If you're not re-thinkng your Dart investment, when Google is, then.......

[Grant Jason]

The killing of the Dart Editor (maybe even Chrome Dev Editor) is in just plain horrible. 

If the company required additional resources to develop the above together with DartPad, then get additional resources - how expensive can that be for Google?

[Grant Jason]

The killing of the Dart Editor (maybe even Chrome Dev Editor) is in just plain horrible. 

If the company required additional resources to develop the above together with DartPad, then get additional resources - how expensive can that be for Google?

On Thursday, May 7, 2015 at 7:44:14 AM UTC+2, David Notik wrote:

[Joel Trottier-Hébert]

@Grant at first I was angry that the VM wouldn't be in Chrome. But at the same time, I knew it from the beginning. It was utopic to think the opposite (at least, I think) at least, for now...

Dart isn't as popular as it should be, and with the rise of typescript, es6, etc, the web isn't the same as when Dart came out (~2011).

I am sure that when Dart gets mature enough, and has a community that's big enough to justify such a change, they'll reconsider it. But that's just me and my dreams :P

Joel Trottier-Hebert

[Jim Trainor]

The editor code is open. Go grab it and continue development.  How much time could it possibly consume?

[Celerity Abbottt]

This is a very insightful, clearly thought out post. Thanks for sharing.

My 2 cents:

- Dart to JS makes more sense (to everyone) on the client side.

- Mobile is hard for Dart to get in.

[Celerity Abbottt]

@Bob: Thanks for sharing and reassurance.

Some questions - if they can be answered:

- What is the UI framework that Google internal Dart projects use the most?

- When can we expect Angular2 Dart catch up with Angular2 JS / Typescript?

- Are the more recent performance benchmarks of Dart vs. JS in whatever application scenarios?

[Günter Zöchbauer]

I haven't tried Angular 2 yet but what makes you think Angular 2 Dart is behind Angular 2 JS?

Angular 2 Dart is autogenerated from Angular 2 TypeScript.

[Anton Moiseev]

Günter, correct. However Angular 2 for Dart requires its own pieces like code transformers for example, hence completeness of TS version doesn't automatically mean feature-parity with Dart version.

BTW, we have a product written with AngularDart and obviously were following ng-conf this year. We were truly impressed by Dave Smith's talk Angular + React = Speed (for those who haven't seen it I highly recommend it), where he showed a blazingly fast grid written with Angular 2. We reproduced exactly his demo but using Angular 2 Dart to check whether Dart version can deliver the same performance. It was noticeably a bit slower than JS version, but still super fast. And we should take into account that it was in very early Angular 2 Dart days + no transformers existed at the moment.

So, we are still happy with Dart after 1.5 year of development, planning migration to Angular 2 later this year and we are going to use Dart for it.

[Filipe Morgado]

Not sure if this is relevant (or even related), but dart2js --trust-type-annotations and --trust-primitives flags do wonders in my experiments.

[Grant Jason]

Thanks for the tip Tim, I'll also just fork Dart whilst I'm at it, since it's also open.

[Celerity Abbottt]

@Filipe  - What kind of wonders did these flags do?

[Günter Zöchbauer]

I think this is where it's explained in detail https://www.youtube.com/watch?v=xneWkyKLuk0

[Filipe Morgado]

From this article.

When comparing Dart and JS performance, people usually compare JS vs dart2js output, both running on V8 in Chrome, rather than Dart VM vs V8.

These flags are not used out of the box.

When I evaluated StageXL as a replacement for Adobe Air, --trust-type-annotations brought a significant performance improvement.

I haven't seriously used Dart for the browser yet, so read this with a grain of salt. 

[kc]

Programming languages usually ride to success on the back of a platform which fulfils a unique need/niche

JS rode the browser, Java the server (via servlets supplanting Perl) serving the browser, Go concurrency with server containers.

Dart imo should have been part of delivering 60fps mobile touch web. Now, Sky looks like the DOM rethought for 60fps+. So given that Sky has gone with Dart, Dart does have the opportunity of a platform to ride on.

I've heard the argument that developers want choices - but not really. Just one modern language which unifies Google's own platforms - Android and Chrome OS - with material design. And can target the server, the broader web, and iOS/Win10.

And in terms of syntax and semantics - pulls together the best of and dumps the worst of:

Java - straightforward - but boilerplate and Oracle.

JS - closures and json - but too weird.

Go - concurrency - but more suited to server systems dev - not the client.

K. 

[Joel Trottier-Hébert]

> I've heard the argument that developers want choices - but not really.

There are tons of cool JS libraries out there. The more we can use with great JS interop, the better the opportunity will be for code reuse. Same thing goes to exported Dart libs as JS.

Joel Trottier-Hebert

[George Moschovitis]

Just one modern language which unifies Google's own platforms...

+1

[Eduardo Teixeira Dias]

One language to rule them all!!!

2015-05-12 15:49 GMT-03:00 George Moschovitis <george.mo...@gmail.com>:

Just one modern language which unifies Google's own platforms...

+1

--

[Celerity Abbottt]

"60Hz Mobile Touch Web" -- sounds really cool.  Patented / trademarked? 

Can only Dart accomplish this?  What about ExtJS Touch or the like?

[Filipe Morgado]

I still use Adobe AIR for everything but the web. It's efficient, productive, has a "DOM-like" Stage, an OpenGL-like Stage3D and targets every major platforms (with AoT compilation for iOs).

In fact, Actionscript 3 was my favorite language for a long time, until I tried Dart. But the Dart platform is still years behind AIR in terms of deployment and UI.

Now comes my uneducated guess about Sky (cut me some slack).

Sky seems to be a modern version of Abode AIR. An optimized DOM subset in its own thread (probably a OpenGL wrapper too), a better concurrency model and services threads written in Dart or other languages. It seems there are bindings for Java and NaCL. So better than a unified language, Sky may be a unified platform for a lot of languages. We all now how people are resistant to change and like to keep using the languages they already know.

Supporting Java bytecode, one could possibly use any JVM-targeted languages. It's a big win.

SInce NaCL is based on LLVM, one could possibly use any language that compiles natively. It's a bigger win.

I'd love to try Rust for performance-critical services.

ART, Dart VM, NaCL and V8 covers the whole world and Sky could be an efficient UI engine with a language-independent message-passing interface that could target any platform.

This goal should have been set years ago.