Dart & AtScript

[kc]

Looks like Angular 2 is unifying dev around a ES6 variant.

K.

[Danny Tuppeny]

Looks like Angular 2 is unifying dev around a ES6 variant.

Oh man, WTF?

A colleague just sent me this doc about AtScript. If this doesn't spell the death of Dart, I don't know what does. Google is investing in another type system for JavaScript to basically solve the sme problems that Dart does; but it's no Dart?

I was just rounding off a Dart prototype in preparation for some tech/framework comparisons next month; but I really don't see how we can pick Dart now. If Google are already creating an alternative to Dart, clearly they don't believe in it.

Really disappointing :O(

[Günter Zöchbauer]

Missing type annotations is not the only thing Dart fixes compared to JavaScript.

This document is an indication where Angular 2.0 is heading to and would rather

discourage me using Angular then discouraging me using Dart.

It will at least take two years until 

[Seth Ladd]

On Fri, Oct 24, 2014 at 7:49 AM, Danny Tuppeny <da...@tuppeny.com> wrote:

Looks like Angular 2 is unifying dev around a ES6 variant.

Oh man, WTF?

A colleague just sent me this doc about AtScript. If this doesn't spell the death of Dart, I don't know what does. Google is investing in another type system for JavaScript to basically solve the sme problems that Dart does;

AtScript does not solve the semantics problem. AtScript is more like a coffeescript. It's a syntax for JavaScript. Dart is it's own language.

 

but it's no Dart?

I was just rounding off a Dart prototype in preparation for some tech/framework comparisons next month; but I really don't see how we can pick Dart now.

Of course you can pick Dart now. It's post 1.0, there are a growing list of packages in pub, a new JS interop mechanism is being developed, async/await on its way, dart2js is getting incremental compilation, there is strong tool support, and an active and supportive community. Dart is real and here now.

 

If Google are already creating an alternative to Dart, clearly they don't believe in it.

I think you're misreading the situation. With all due respect, you're writing FUD.

There are many teams in Google, and each picks a set of technologies that work for them (GWT, Dart, ES6, Traceur, Closure compiler, etc). Many teams are innovating, which is great.

dart2js, Dartium, dart:html continue.

 

Really disappointing :O(

--
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.

[Danny Tuppeny]

On 24 October 2014 16:13, Günter Zöchbauer <gzo...@gmail.com> wrote:

Missing type annotations is not the only thing Dart fixes compared to JavaScript.

The problem is not what AtScript is/isn't, but that teams within Google are actively choosing not to use Dart for new projects; even those that already have Dart versions!

I don't know if this is down to internal politics or something they know that we don't; but unless Google start being transparent about what's going on; there's no way I can push Dart for use in our commercial apps.

[Kasper Lund]

Danny,

Google is investing in a lot of things, but the Dart team is committed
to making Dart a fantastic choice for developers. A part of the reason
why we're doing Dart is that we feel that 'language monoculture' is
hurting innovation and progress -- so in that respect we definitely
welcome new initiatives, including new languages. Personally, I think
Dart stacks up pretty nicely against most (all?) of the alternatives,
but I guess I'm somewhat biased :-)

Cheers,
Kasper

[Danny Tuppeny]

On 24 October 2014 16:15, 'Seth Ladd' via Dart Misc <mi...@dartlang.org> wrote:

I think you're misreading the situation. With all due respect, you're writing FUD.

Google aren't exactly being transparent about what's going on. There were already questions recently when the Angular Team announced they weren't using Dart, and they went unanswered.

If we're drawing the wrong conclusions, it's because of a lack of information. If Google won't talk about these things, or the relationships between them, we're going to get nervous. The Dart VM is still not in Chrome.

I'm not trying to start an argument; I am genuinely disappointed that any team within Google would rather work on something new than help Dart solve the needs they have. There is nothing more that I want than to use Dart instead of JavaScript; but if we go all-in on Dart, it's hard to turn back (the generated JS is not usable). I don't want all this on my shoulders if Dart collapses.

If Google had launched some of their own apps in Dart; it would be easier to believe they were invested in Dart; but even Inbox didn't use it :(

[Danny Tuppeny]

On 24 October 2014 16:17, 'Kasper Lund' via Dart Misc <mi...@dartlang.org> wrote:

Personally, I think
Dart stacks up pretty nicely against most (all?) of the alternatives,
but I guess I'm somewhat biased :-)

I think so too; which is why I was surprised that the Angular Team are choosing not to use it (and to invest so heavily in re-implementing parts of it). Without clear answers to why, how can we be anything other than paranoid? Maybe the reasons they're choosing not to use it apply to us to? Maybe they know Dart VM is never coming to Chrome?

If information is not forthcoming, we have to assume the worst. How can we be confident enough to write new codebases that need to live for 10+ years in a language that teams within your own company appear to be going to great lengths not to use? :O(

[Seth Ladd]

Inbox was started before Dart was mature.

[Justin Fagnani]

Please consider that Angular is a project with a very large JavaScript user base. They were rightfully never going to abandon that user base, and so we're always going to keep a JavaScript version in development. Remember that Angular dropping JavaScript support was a huge fear from their community when they announced the Dart port and Angular was unequivocal about continuing to support JavaScript.

You can take AtScript as a sign that Angular liked Dart enough that they wanted some of the advantages in JavaScript for Angular 2.0.

--

[Günter Zöchbauer]

On Friday, October 24, 2014 5:27:42 PM UTC+2, Danny Tuppeny wrote:

On 24 October 2014 16:17, 'Kasper Lund' via Dart Misc <mi...@dartlang.org> wrote:

Personally, I think
Dart stacks up pretty nicely against most (all?) of the alternatives,
but I guess I'm somewhat biased :-)

I think so too; which is why I was surprised that the Angular Team are choosing not to use it (and to invest so heavily in re-implementing parts of it). Without clear answers to why, how can we be anything other than paranoid? Maybe the reasons they're choosing not to use it apply to us to? Maybe they know Dart VM is never coming to Chrome?

One reason could be that Angular.js has gained a pretty large user base but the current Angular.dart user base is tiny in comparsion and they don't want to develop in two different languages. To keep the gained users happy they might consider it a higher priority to provide a solution that works for JavaScript users instead of trying to lure them to Dart and might follow a similar strategy like Polymer.dart to build a wrapper around JavaScript libraries that provides some comfort for Angular.dart users too.

The main differences are 

that the JavaScript parts of Polymer are only (mostly?) the polyfills for features not yet supported by browsers natively but very probably will be in the future.

and Angular built with AtScript will not profit from native Dart support in Chrome.

[Filipe Morgado]

I wouldn't worry much about it.

As Seth said, a lot of exciting features are coming to Dart. And I can imagine a few more for 2.0 (non-nullable types wouldn't break anything and could prevent lots of noSuchMethod).

Dart is here to stay, with or without Chrome.

I'm loving Dart on the server.

About Chrome ...

Oilpan is going on nicely, I think there's been Chromium guys committing to Dartium (something about GC inter operation?) and the Dart VM is getting fine-grained optimizations (such as jitted RegEx).

I'd say the multi-vm thing is happening anyway and Dart is getting a front-row seat.

IMO, the Angular guys are doing a bad move.

They're starting from scratch what has already been done 50 times (CoffeeScript, TypeScript, MyScript, YourScript, etc)

I don't know where they're going, but I doubt AngularDart is going away.

[Matthew Butler]

w

On Friday, October 24, 2014 12:15:46 PM UTC-3, Danny Tuppeny wrote:

On 24 October 2014 16:13, Günter Zöchbauer <gzo...@gmail.com> wrote:

Missing type annotations is not the only thing Dart fixes compared to JavaScript.

The problem is not what AtScript is/isn't, but that teams within Google are actively choosing not to use Dart for new projects; even those that already have Dart versions!

No actually in this case for Angular the problem was that they were maintaining two similar but slightly different entire versions of their framework. From my understanding the core of Angular 2.0 will be written in AtScript, at which point the JavaScript and Dart versions of Angular will be able to share significant codebase without the need to duplicate the entire framework. This is already being done in Polymer/Polymer.dart (including core/paper elements) why is it an issue for Angular to do the same?

Unfortunately Dart is not yet ready for 'compile and use as a JavaScript library' from my understanding. The steps are in that direction as well but ultimately it's much easier to include a Javascript library in Dart than vice versa.

 

I don't know if this is down to internal politics or something they know that we don't; but unless Google start being transparent about what's going on; there's no way I can push Dart for use in our commercial apps.

I think Google has been extremely transparent with Dart. As mentioned very recently it was used with the Polymer libraries for the Google Elections site (again). Work is still heavily continuing on the Oilpan project with numerous contributors from Google and external (such as opera). I've followed the work and am continually astounded on how much work is going into this project and how much still appears to be left. It gave me a grasp of the blink codebase that I never even conceived of before. 

Simply because Angular wants to create a common code base doesn't mean all the work that the Dart team has done, and continues to do (along with other areas like oilpan) is all going to be brushed aside.

Matt

(Note: I'm not a Google employee and don't have any internal view of matters, only basing my information on information that's publicly available such as the tracking bug on the chromium project: https://code.google.com/p/chromium/issues/detail?id=340522, Blink and Dart groups etc). 

[George Moschovitis]

This AtScript thing is really annoying. If they just wanted JavaScript with types they could have just used TypeScript.

-g.

[Warren]

Indeed. 

I can't see Angular JS users being happy with yet another language. 

[Istvan Soos]

AFAIK the point of AtScript is not to suggest that Dart or JS
developers should use it, rather it provides a framework for the
Angular core team to develop Angular in a way that will satisfy both
audiences. I think one of the reasons to do this was to adapt the
Java/Dart-like annotation metadata, which was not present in
TypeScript (I may be wrong on this one, don't hold me to it). Creating
something new and having in full control of it also helps to go for a
brand new version.

You can already look at the code an try it out. It generates both JS
and Dart version of the Angular modules. It translates both the code
to tests, and can run them. It is far from being ready, but you can
get the idea where they are going:
https://github.com/angular/angular

In Google, teams have freedom to choose whatever they see fit, and
their team has chosen to create AtScript. Looking at Google teams what
they have selected is a not an indicator in any way, and you should
always do your own judgement and evaluation.

Istvan

[Cogman]

Yah, I don't see a good reason for the existence of atscript (other than inventing a language just because you can).

If anything, I think they should have pushed for better a better dart library/js library interop story and made angular dart "Angular 2.0".  That would be a huge boon for dart and the angular dart library.

[kc]

AtScript isn't so much a new language as:

ES6
+ TypeScript-ish ':' type annotations
+ '@' annotations

Seems like the AtScript and TypeScript teams aim to collaborate to get these annotation ideas into post-ES6. Its a reasonable strategy.

K.

[Göran Johansson]

What if Angular is the past? And Dart the future...

[Daniel Davidson]

I think an important point is Angular has a different audience than Inbox. Angular is writing frameworks for developers. Apparently they have a very large javascript client base and a hopefully growing dart client base. When the audience is tech consumers using your tools to create applications you might behave differently than when your audience is for the general end user. The reason this might not be much of a concern in terms of canabalization is that the vast majority choosing between Dart and JavaScript are trying to build something for the end user and want the best language, tools, environment, experience.

When I first read this thread I thought - oh boy another NIH syndrome. I don't work there, but I'll bet it is a mischaracterization to say "team within Google would rather work on something new than help Dart solve the needs they have". I bet if the team is reasonably big there will be developers there looking at and being required to use AtScript who would probably want nothing more than to drop JavaScript and use all the goodies in Dart. I don't know or use JavaScript at all (thanks to Dart), but I am now thankful they are taking this approach because to do otherwise would leave the original Angular JS customers in the lurch. It is actually very customer focused and user friendly behavior. 

[Thomas Schranz]

As I see it the Angular team plans to provide a JS and a Dart version of Angular going forward.

This is great for users of Angular (whether you are using JS or Dart).

For core developers maintaining a project for multiple platforms always is a challenge though.

Instead of building the core part of the library using a lowest common denominator (ES5?)

they wanted to use something more powerful and their approach was using (Traceur to get JS.next plus some custom additions they wanted to see).

AtScript as a solution for this challenge might look complicated or elegant depending from where you come from.

As I see it the Dart project (in terms of vision/aspiration) goes way beyond AtScript.

Dart is a platform (language, runtime(s), packages, ides, …)

That said, I can see how at first glance the Dart and AtScript efforts might look similar.

Nothing to worry about though, looking fwd to Angular.dart

[Thomas Schranz]

To adress & repeat the 'Google' concerns: I'm not part of Google and never was so

you might take this with more or fewer grains of salt:

The company is so huge that coordination happens using OKRs (objectives and key results)

this scales incredibly well in terms of aligning people in terms of what really matters (in which direction to go and why).

The fascinating thing about OKRs though is that people derive the 'how' from the 'why' which enables autonomy and efficiency

but at the same time (typical trade-off) it might result in similar efforts (way better than people not being aligned and doing whatever).

From a bird's eye view this is can be a great trade-off as the company is not crawling to a halt as it grows (compare Google to other companies)

and it might yield solutions that overlap which isn't inherently bad either. Over time consolidation usually happens anyway, it just takes time.

I also fell into the trap of analyzing Google's (or Apple's or Mozilla's) behaviour as if it was a company of 2-3 people

that make every decision. When I realized they don't function like that but more like a system based on certain dynamics

everything made a lot more sense to me.

My 2 cents as a non-Googler: Nothing to worry about.

[kc]

MS is getting it's platform and development act together after a decade of mismanagement.

Google needs more joined up thinking. ART is in ChromeOS but not Dart.

K.

[George Moschovitis]

MS is getting it's platform and development act together after a decade of mismanagement.

Google needs more joined up thinking. ART is in ChromeOS but not Dart.

+1

[Danny Tuppeny]

On 27 October 2014 19:51, kc <kevin...@gmail.com> wrote:

Google needs more joined up thinking. ART is in ChromeOS but not Dart.

After a little more thinking; I think it's the Angular Team who are mad, not Google. I blogged my thoughts in another OTT, click-bait-y, doesn't-have-all-the-facts blog post here:

http://blog.dantup.com/2014/10/have-the-angular-team-lost-their-marbles/

;O)

[Warren]

I am genuinely curious to know why the Angular team could not have invested in improvements in dart2js rather than invent a new language (there are probable good reasons, but I'd like to hear a little of the rationale). 

Even forking dart2js to produce a special "dart2angularjs" would seem preferable. To Danny's point - AtScript is going to require new tooling and editor support, which seems like a lot of effort for a niche language.

I am also curious as how AtScript will impact the edit/compile/debug cycle.  Does AtScript introduce a new compile step into the cycle, or is it a one time penalty?

[Günter Zöchbauer]

This is a very lame blog post and mostly uninformed ranting.

Most people always want perfect support for all the newest features and tech but rant about changes.

I'm also not very happy about the direction Angular is taking but there are very smart people working on Angular and I can imagine, that I just haven't enough knowledge to understand the decisions.

[Danny Tuppeny]

On 29 October 2014 15:19, Günter Zöchbauer <gzo...@gmail.com> wrote:

This is a very lame blog post and mostly uninformed ranting.

My blog is for my thoughts; which are almost always lame and uninformed ranting. Though I'd actually say this post is more informed than most; I'd love to know which bits you think are uninformed, so I can update them. As far as I can tell, the new syntax, the estimated timeline and the estimated EOL of 1.3 are all based on information from the Angular team.

 

Most people always want perfect support for all the newest features and tech but rant about changes.

Ofcourse they do; but that doesn't mean they should get them. Authors of tools and frameworks need to push back and focus on stability; some of us are trying to focus on writing software, not rewriting syntax every few years.

 

I'm also not very happy about the direction Angular is taking but there are very smart people working on Angular and I can imagine, that I just haven't enough knowledge to understand the decisions.

I don't doubt that the decisions have been thought about. I just suspect that the value of "doing something shiny and new" and "breaking everything" have different weight to the majority of software developers (that aren't based in Silicon Valley). I'm simply providing my thoughts and feedback; people are free to disagree and disregard them; they're only my thoughts. If everyone sits silently, how to framework authors know if what they're doing is perceived to be good or bad?

[Hans - Jürgen Alps]

+1 :)

--

[Günter Zöchbauer]

On Wednesday, October 29, 2014 5:55:05 PM UTC+1, Danny Tuppeny wrote:

On 29 October 2014 15:19, Günter Zöchbauer <gzo...@gmail.com> wrote:

This is a very lame blog post and mostly uninformed ranting.

My blog is for my thoughts; which are almost always lame and uninformed ranting. Though I'd actually say this post is more informed than most; I'd love to know which bits you think are uninformed, so I can update them. As far as I can tell, the new syntax, the estimated timeline and the estimated EOL of 1.3 are all based on information from the Angular team.

Sorry, if I knew you did the post I would have written it a bit differently ;-)

I had the impression you just linked a blog post you found.

Especially the rant about the new syntax. There were severe limitations with the old syntax and they published a conceptual document a few months ago where the proposed changes and the reasons were discussed in public. A main point was the support for web components which is quite important.

 

 

Most people always want perfect support for all the newest features and tech but rant about changes.

Ofcourse they do; but that doesn't mean they should get them. Authors of tools and frameworks need to push back and focus on stability; some of us are trying to focus on writing software, not rewriting syntax every few years.

I see this quite differently. 

If something is broken, fix it. The moment you stop fixing things because you want to avoid breaking changes the framework/language/... starts dying (with exponential speed).

Which leads just to other rants from developers who invested so much time/effort/money in a framework and suddenly have more and more troubles using the framework with the new hotness in web/mobile/cloud-development.

If one develops a language or framework the most important skill seems to be to ignore ranting from devs who fear and try to avoid change (which is unavoidable anyway).

 

I'm also not very happy about the direction Angular is taking but there are very smart people working on Angular and I can imagine, that I just haven't enough knowledge to understand the decisions.

I don't doubt that the decisions have been thought about. I just suspect that the value of "doing something shiny and new" and "breaking everything" have different weight to the majority of software developers (that aren't based in Silicon Valley). I'm simply providing my thoughts and feedback; people are free to disagree and disregard them; they're only my thoughts. If everyone sits silently, how to framework authors know if what they're doing is perceived to be good or bad?

I don't use Angular a lot but I follow the discussions in the GitHub issues and Google groups and I got a quite different impression.

 

[Danny Tuppeny]

My blog is for my thoughts; which are almost always lame and uninformed ranting. Though I'd actually say this post is more informed than most; I'd love to know which bits you think are uninformed, so I can update them. As far as I can tell, the new syntax, the estimated timeline and the estimated EOL of 1.3 are all based on information from the Angular team.

Sorry, if I knew you did the post I would have written it a bit differently ;-)

I had the impression you just linked a blog post you found.

I should've put my name in the domain ;-)

But you shouldn't word things differently; I appreciate all the feedback; however it's written. I learned almost everything I know from being wrong. Repeatedly ;)

 

Especially the rant about the new syntax. There were severe limitations with the old syntax and they published a conceptual document a few months ago where the proposed changes and the reasons were discussed in public. A main point was the support for web components which is quite important.

I understand there are reasons; but I just don't agree the new syntax is a good way to go. There were already weird "magic strings" in Angular that I had to look up every time I used them; but now we have more non-discoverable conventions. I would rather they were obvious to look at, than a random collection of symbols assigned for each purpose. There's enough cryptic stuff in programming, we don't need more. We're building a world around technology that only a tiny portion of the population can understand/write; we need to be making it more obvious; or we're setting ourselves up to fail! (though, I guess in my lifetime, it will continue to have "positive effects" on our salaries ;-))

Ofcourse they do; but that doesn't mean they should get them. Authors of tools and frameworks need to push back and focus on stability; some of us are trying to focus on writing software, not rewriting syntax every few years.

I see this quite differently. 

If something is broken, fix it. The moment you stop fixing things because you want to avoid breaking changes the framework/language/... starts dying (with exponential speed).

Which leads just to other rants from developers who invested so much time/effort/money in a framework and suddenly have more and more troubles using the framework with the new hotness in web/mobile/cloud-development.

If one develops a language or framework the most important skill seems to be to ignore ranting from devs who fear and try to avoid change (which is unavoidable anyway).

There are times to make huge sweeping breaking changes, and that is before you hit v1. People are heavily invested in Angular, and it's irresponsible to discount all of their effort with no consultation. If thee really is no other way to progress with v2, then time should be invested in building tools to aid the migration; or a way that it can be done progressively. We shouldn't have to down tools and do no releases for 6 months while we rewrite all our Angular code just to be on a supported version; we should've been able to do it a little with each release. Even if this was possible, we'd then end up with an enormous release at the end which is high risk. What happened to deploying continuously?

I'm not saying don't break things. Just break them in a slow, careful, progressive way, and/or provide tools to help your users. It feels like all of the decisions where made by the Angular Team and not by existing Angular users that have already invested hugely in their codebases.

 

I don't doubt that the decisions have been thought about. I just suspect that the value of "doing something shiny and new" and "breaking everything" have different weight to the majority of software developers (that aren't based in Silicon Valley). I'm simply providing my thoughts and feedback; people are free to disagree and disregard them; they're only my thoughts. If everyone sits silently, how to framework authors know if what they're doing is perceived to be good or bad?

I don't use Angular a lot but I follow the discussions in the GitHub issues and Google groups and I got a quite different impression.

I spent weeks prototyping in Angular many months ago (before I changed to Dart) in preparation for discussions about our tech stack. As I got more into Dart (and grew to hate Angular), one of my (very experienced) colleagues took over the job of an Angular prototype instead. He started clean, to not be influenced by my prototype. He also grew to hate it (for mostly for similar reasons to be, which are documented in this post). These changes just add fuel to the fire.

I'm not saying we're more typical devs than the Angular fans; we may be in a minority; but if I was the Angular Team; I would love to hear from people that are having trouble with it; so it can feed into future decisions (or at least consciously be discarded). If I authored a JavaScript framework (don't worry, I never will), I would care more about what the haters say then the fans. So, I will continue to blog when I dislike something; it's how I would "want to be treated". I've learned far more from discussions and debates when I disagree with people than from any number of books or training courses.

(I feel we're somewhat veering off-topic, but I'd encourage commenting on the original post if you have something to say :))

[kc]

AtScript isn't a new language - it's ES6. TC39 tentatively agreed Typescript type annotation syntax.

So Angular are betting on browsers natively implementing ES6 and Web Components and building on it.

K.