[dart-misc] Dart & Sky - Performance

[kc]

And info available on how an Android app coded in Dart/Sky compares to a
'native' Java/XML in terms of performance and resource usage
(mem/cpu/battery) etc.

I bang on about syntax/semantics but if performance is ballpark to 'native'
then Dart could be very successful regardless (and give a catalyst to
dart2js). The dev experience will be so much nicer. (Though fix the
Java-ism's pls).

Dart was outflanked by ES6/TS on the web but could succeed via Sky/mobile.

K.

[kc]

Re tweaks.

Here's some Dart/Sky code:
https://github.com/domokit/mojo/blob/master/examples/dart/camera_roll/camera_roll.dart

Picture draw(Image image) {
PictureRecorder canvas = new PictureRecorder(view.width, view.height);
Paint paint = new Paint()..color = const Color.fromARGB(255, 0, 255, 0);
canvas.scale(view.width / image.width, view.height / image.height);
canvas.drawImage(image, 0.0, 0.0, paint);
return canvas.endRecording();
}

First contact - Java! - and:
1) Type inference? new?
2) What's the range operator doing. Cascasde? Does '.' bind less then '..'?
3) const? Performance or semantics?

Tweaked:

Picture draw(Image image) {
let canvas = PictureRecorder(view.width, view.height);
let paint = Paint().color = Color.fromARGB(255, 0, 255, 0);
canvas.scale(view.width / image.width, view.height / image.height);
canvas.drawImage(image, 0.0, 0.0, paint);
return canvas.endRecording();
}


Not so much ceremony and boilerplate. And if Checked Mode is removed from
the VM then type annotations (whatever form they take) can evolve.
Trivial - but gives a different vibe.

Also, regarding fresh thinking re concurrency, 'blocks', non-local returns:

This async is tricky to eyeball imo:

void drawNextPhoto() {
var future = cameraRoll.ptr.getPhoto(photoIndex);
future.then((response) {
if (response.photo == null) {
cameraRoll.ptr.update();
photoIndex = 0;
drawNextPhoto();
return;
}
}
}


Sure await/async sugar is good. But with some form of 'block' or trailing
closure:

void drawNextPhoto() {
cameraRoll.ptr.getPhoto(photoIndex) do { response ->
if (response.photo == null) {
cameraRoll.ptr.update();
photoIndex = 0;
drawNextPhoto();
return;
}
}
}

Expresses intent and maybe a decent fit for where concurrency is going?

Other semantics like value objects/immutability, pattens, dependency
injection are very welcome - but really only if they add value to the
modern mobile performance and dev experience.
For instance would value objects/immutability baked-in benefit both Sky and
Angular - as well as allowing dev's to express their intent more clearly.

K.

Any info available on how an Android app coded in Dart/Sky compares to
'native' Java/XML in terms of performance and resource usage
(mem/cpu/battery) etc.
I bang on about syntax/semantics but if performance is ballpark to
'native' then Dart could be very successful regardless (and give a catalyst
to dart2js). The dev experience will be so much nicer. (Though fix the
Java-ism's pls).
Dart was outflanked by ES6/TS on the web but could succeed via Sky/mobile.
K.

[Jos Hirth]

Post by kc
2) What's the range operator doing. Cascasde? Does '.' bind less then '..'?

Paint paint = new Paint()..color = const Color.fromARGB(255, 0, 255, 0);

Is the same as:

Paint paint = new Paint();
paint.color = const Color.fromARGB(255, 0, 255, 0);

https://www.dartlang.org/docs/dart-up-and-running/ch02.html#classes

[Filipe Morgado]

I would add that it becomes even more awesome when you need to set multiple
properties:

final element = new DivElement()
..id = 'errorId'
..className = 'errorStyle'
..text = 'Some Message'
..title = 'Some Title';

[kc]

I'm familiar with the cascade operator. It's useful.

The point was that with Flutter/Sky an entire *new *audience could be
looking at Dart. How does the language look to them at first
glance/contact. '..' is the 'familiar' range operator for most people.

Currently Dart looks a little too much like Java with some additional
strange bits on top.

K.

Post by Filipe Morgado
I would add that it becomes even more awesome when you need to set
final element = new DivElement()
..id = 'errorId'
..className = 'errorStyle'
..text = 'Some Message'
..title = 'Some Title';

[kc]

Lets hope Dart 2 isn't a Perl 6.

K.

On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need to set
multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777
0095
http://www.stonehenge.com/merlyn/>
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig

[Eduardo Teixeira Dias]

Short list for Dart 2:
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic

The list of failures examples is huge:
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn

- JBoss Seam
- Incompatibility: Seam 2 / Seam 3

- AngularDart
- Obsolete from the start

- Python
- Incompatibility


Learn from history:
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/

Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need to
set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777
0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--
Saudações,

Eduardo Teixeira Dias

------------------------------------------
Tendencies Consultoria Ltda.
Tel: [phone]
Cel: 11 9 9246-4192

[Gen]

I have no Dart code to lose or to port.
So "100% backward compatibility" is certainly not on my list.
The standard library is already statically and strongly typed because that
is the main reason for using Dart and not Javascript.
Only the compiler must change to support Dart 2.
With Dart 2, the compiler and build tools could use a language version
property declared somewhere. If it is absent then treat the code as Dart 1
code.

I do not remember having read something against Python 3 and the need for
breaking changes.
The link you gave mentions that only 13 % think that Python 3 was a mistake
and maybe some of these people think that Python 3 did not change enough.
AFAIK, the Python 2to3 tool works well.

Besides, I wonder why List<int> and not ByteBuffer is used in dart:io.
Maybe it does not matter for performance, does it ?

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn
- JBoss Seam
- Incompatibility: Seam 2 / Seam 3
- AngularDart
- Obsolete from the start
- Python
- Incompatibility
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/
<http://www.google.com/url?q=http%3A%2F%2Fwww.randalolson.com%2F2015%2F01%2F30%2Fpython-usage-survey-2014%2F&sa=D&sntz=1&usg=AFQjCNGGpdDg-MVf0jn374_zuzOKyZXNoA>
Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need to
set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777
0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--
Saudações,
Eduardo Teixeira Dias
------------------------------------------
Tendencies Consultoria Ltda.
Tel: [phone]
Cel: 11 9 9246-4192

[Eduardo Teixeira Dias]

100% backward compatibility shoud be a goal.

It shows the entertprise that Dart is safe for investment.

I have Java code with more than 10 years that runs fine on current Java VMs.

For me, being able to to write modern Rich Client Web Apps, Mobile Apps and
Server APIs (Redstone) using the same language is amazing but backward
compatibility is essential to convince customers.

Customers like that.

Performance gains and new functionalities are always wellcome!

For companies with large Python 2 code bases it is problem not necessarily
for the developers...

Post by Gen
I have no Dart code to lose or to port.
So "100% backward compatibility" is certainly not on my list.

So, you wrote no Dart code worth porting, cool...
The standard library is already statically and strongly typed because that

Post by Gen
is the main reason for using Dart and not Javascript.
Only the compiler must change to support Dart 2.
With Dart 2, the compiler and build tools could use a language version
property declared somewhere. If it is absent then treat the code as Dart 1
code.
I do not remember having read something against Python 3 and the need for
breaking changes.
The link you gave mentions that only 13 % think that Python 3 was a
mistake and maybe some of these people think that Python 3 did not change
enough.
AFAIK, the Python 2to3 tool works well.
Besides, I wonder why List<int> and not ByteBuffer is used in dart:io.
Maybe it does not matter for performance, does it ?

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn
- JBoss Seam
- Incompatibility: Seam 2 / Seam 3
- AngularDart
- Obsolete from the start
- Python
- Incompatibility
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/
<http://www.google.com/url?q=http%3A%2F%2Fwww.randalolson.com%2F2015%2F01%2F30%2Fpython-usage-survey-2014%2F&sa=D&sntz=1&usg=AFQjCNGGpdDg-MVf0jn374_zuzOKyZXNoA>
Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need to
set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777
0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--

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

['Bob Nystrom' via Dart Misc]

On Thu, Oct 8, 2015 at 9:57 AM, Eduardo Teixeira Dias <

Post by Eduardo Teixeira Dias
I have Java code with more than 10 years that runs fine on current Java VMs.

But do you have any Oak
<https://en.wikipedia.org/wiki/Oak_(programming_language)> code that you
still use? CFront or CPre <https://en.wikipedia.org/wiki/Cfront>? ABC
<https://en.wikipedia.org/wiki/ABC_(programming_language)>?

I think the interesting question is is Dart more like Java 1.0 right now,
or more like Oak? Obviously, nominally, it's more like Java 1.0. We have
publicly launched, declared it 1.0, have an open spec, etc.

In terms of external usage, I don't know if I'd say we're successful enough
to really be beholden to our initial version.

*Certainly* once we reach a significant level of mindshare, compatibility
is absolutely paramount. No one disputes that. But every new product has a
period of instability while it figures out how to fit itself to the market
before it does that. I'm not sure if Dart has realistically reached that
threshold yet.

– bob

[Eduardo Teixeira Dias]

My point is: Dart will NEVER get mindshare if is not persieved by managers
as safe for investment.

Dart is more than good enough to take off. It needs more credibility,
tools and documentation.

Credibility come from actions. Been credible. Not promising that in the
next version will be credible. Show that it is safe, been safe.

The aleternative is. Continue to throw incompatible alphaware after
incompatible alphaware hopping that managers some how will someday believe
that this new version will be safe to invest in.

Post by 'Bob Nystrom' via Dart Misc
On Thu, Oct 8, 2015 at 9:57 AM, Eduardo Teixeira Dias <

Post by Eduardo Teixeira Dias
I have Java code with more than 10 years that runs fine on current Java VMs.

But do you have any Oak
<https://en.wikipedia.org/wiki/Oak_(programming_language)> code that you
still use? CFront or CPre <https://en.wikipedia.org/wiki/Cfront>? ABC
<https://en.wikipedia.org/wiki/ABC_(programming_language)>?
I think the interesting question is is Dart more like Java 1.0 right now,
or more like Oak? Obviously, nominally, it's more like Java 1.0. We have
publicly launched, declared it 1.0, have an open spec, etc.
In terms of external usage, I don't know if I'd say we're successful
enough to really be beholden to our initial version.
*Certainly* once we reach a significant level of mindshare, compatibility
is absolutely paramount. No one disputes that. But every new product has a
period of instability while it figures out how to fit itself to the market
before it does that. I'm not sure if Dart has realistically reached that
threshold yet.
– bob
--
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

[kc]

On Thursday, October 8, 2015 at 6:53:49 PM UTC+1, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
My point is: Dart will NEVER get mindshare if is not persieved by managers
as safe for investment.
Dart is more than good enough to take off. It needs more credibility,
tools and documentation.
Credibility come from actions. Been credible. Not promising that in the
next version will be credible. Show that it is safe, been safe.

True. The action which did not take place was the Dart VM going into Chrome
which cost this project credibility.

Post by Eduardo Teixeira Dias
The aleternative is. Continue to throw incompatible alphaware after
incompatible alphaware hopping that managers some how will someday believe
that this new version will be safe to invest in.

No. The alternative is to make some judicious fixes to the language so
that's it very appealing to the people who will check out Flutter. That
Dart and Flutter are a far more expressive and performant way of coding
mobile apps compared to the competition.

The investment of Dart 1.x will be preserved by a fix script from 1.x to
2.x - like with Swift 1.x to 2.x.

K

Post by Eduardo Teixeira Dias

Post by 'Bob Nystrom' via Dart Misc
On Thu, Oct 8, 2015 at 9:57 AM, Eduardo Teixeira Dias <

Post by Eduardo Teixeira Dias
I have Java code with more than 10 years that runs fine on current Java VMs.

But do you have any Oak
<https://en.wikipedia.org/wiki/Oak_(programming_language)> code that you
still use? CFront or CPre <https://en.wikipedia.org/wiki/Cfront>? ABC
<https://en.wikipedia.org/wiki/ABC_(programming_language)>?
I think the interesting question is is Dart more like Java 1.0 right now,
or more like Oak? Obviously, nominally, it's more like Java 1.0. We have
publicly launched, declared it 1.0, have an open spec, etc.
In terms of external usage, I don't know if I'd say we're successful
enough to really be beholden to our initial version.
*Certainly* once we reach a significant level of mindshare,
compatibility is absolutely paramount. No one disputes that. But every new
product has a period of instability while it figures out how to fit itself
to the market before it does that. I'm not sure if Dart has realistically
reached that threshold yet.
– bob
--
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

[Benjamin Strauß]

Post by Eduardo Teixeira Dias
The aleternative is. Continue to throw incompatible alphaware after

Post by Eduardo Teixeira Dias
incompatible alphaware hopping that managers some how will someday believe
that this new version will be safe to invest in.

No. The alternative is to make some judicious fixes to the language so
that's it very appealing to the people who will check out Flutter. That
Dart and Flutter are a far more expressive and performant way of coding
mobile apps compared to the competition.

I agree. I switched from Dart to TypeScript for frontend development a
while ago.

At this point my only interest in Dart is the VM. Namely server side
development and sky/flutter.

Dart should cut off old habits in favor for a brighter future in those
areas.

[kc]

Post by Benjamin Strauß

Post by Eduardo Teixeira Dias
The aleternative is. Continue to throw incompatible alphaware after

Post by Eduardo Teixeira Dias
incompatible alphaware hopping that managers some how will someday believe
that this new version will be safe to invest in.

No. The alternative is to make some judicious fixes to the language so
that's it very appealing to the people who will check out Flutter. That
Dart and Flutter are a far more expressive and performant way of coding
mobile apps compared to the competition.

I agree. I switched from Dart to TypeScript for frontend development a
while ago.
At this point my only interest in Dart is the VM. Namely server side
development and sky/flutter.
Dart should cut off old habits in favor for a brighter future in those
areas.

Given Edge and the fact that the DOM story is becoming clearer maybe Dart
2.0 could target modern browsers exclusively.

K.

[Günter Zöchbauer]

Dart only supports modern browsers.

Post by kc

Post by Benjamin Strauß

Post by Eduardo Teixeira Dias
The aleternative is. Continue to throw incompatible alphaware after

Post by Eduardo Teixeira Dias
incompatible alphaware hopping that managers some how will someday believe
that this new version will be safe to invest in.

No. The alternative is to make some judicious fixes to the language so
that's it very appealing to the people who will check out Flutter. That
Dart and Flutter are a far more expressive and performant way of coding
mobile apps compared to the competition.

I agree. I switched from Dart to TypeScript for frontend development a
while ago.
At this point my only interest in Dart is the VM. Namely server side
development and sky/flutter.
Dart should cut off old habits in favor for a brighter future in those
areas.

Given Edge and the fact that the DOM story is becoming clearer maybe Dart
2.0 could target modern browsers exclusively.
K.

[kc]

Post by Günter Zöchbauer
Dart only supports modern browsers.

Modern != IE.
Modern >= Safari 8

K.

[Matthew Butler]

On Thursday, October 8, 2015 at 1:57:32 PM UTC-3, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.
Yes but the whole point of the discussion is for Dart 2.0. If it were to

remain 100% backward compatible it would be a minor version update, not a
major version number change.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
As of Java 6 (I'm not sure about later, I haven't looked), java internal

version was still 1.6.x (eg everything from Java 2 and up were only minor
version rolls, not a major version. This may have change since that time)

[Eduardo Teixeira Dias]

If you go incremental you do not lose packages.

The programers time are used to do apps and not to write the basics all
over again...

The cost of fragmentation is high.

You want some package that only exists for Dart 1 but you also need a
package that only exists in Dart 2.

As you see the benefits of the incompatibility is enourmous...

It is important to look also if it is not possible, in some way, get the
benefit of the change in a compatible incremental way.

If you just say "We are considering an incompatible Dart 2" all managers
that I know of will wait to see what monster will that Dart 2 look like
before even considering investing in Dart 1.

If in the other hand you say "We are considering an amazing new Dart 2 that
is backward compatiple". You know, is different. I can invest in Dart 1 now.

You can only look at AngularDart forum. The day the Angular announced the
future incompatible Angular2, no upgrade path. The conversations just
stoped. Why??? Let's see what this angular2 will look like from a Dart
developer point of view.

For the record, Angular 2 is looking great.

Post by Matthew Butler
On Thursday, October 8, 2015 at 1:57:32 PM UTC-3, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.
Yes but the whole point of the discussion is for Dart 2.0. If it were to

remain 100% backward compatible it would be a minor version update, not a
major version number change.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
As of Java 6 (I'm not sure about later, I haven't looked), java internal

version was still 1.6.x (eg everything from Java 2 and up were only minor
version rolls, not a major version. This may have change since that time)
--
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

--
Saudações,

Eduardo Teixeira Dias

------------------------------------------
Tendencies Consultoria Ltda.
Tel: [phone]
Cel: 11 9 9246-4192

[Gen]

Do you fear any change in particular ?

I see potential for incompatible changes only with regard to the type
system when:
- "var" will stand for static type interference.
- "strong mode" is the new standard. This might break much in your code. It
remains to be seen how much breaks in the standard library and for other
people.
How much code that is rejected in strong mode is also bad code anyway ?

I agree with you that losing existing code is no option.
But I bet that the Dart team will not confront you with that case unless
there is a compelling reason which I do not see.

Post by Eduardo Teixeira Dias
If you go incremental you do not lose packages.
The programers time are used to do apps and not to write the basics all
over again...
The cost of fragmentation is high.
You want some package that only exists for Dart 1 but you also need a
package that only exists in Dart 2.
As you see the benefits of the incompatibility is enourmous...
It is important to look also if it is not possible, in some way, get the
benefit of the change in a compatible incremental way.
If you just say "We are considering an incompatible Dart 2" all managers
that I know of will wait to see what monster will that Dart 2 look like
before even considering investing in Dart 1.
If in the other hand you say "We are considering an amazing new Dart 2
that is backward compatiple". You know, is different. I can invest in Dart
1 now.
You can only look at AngularDart forum. The day the Angular announced the
future incompatible Angular2, no upgrade path. The conversations just
stoped. Why??? Let's see what this angular2 will look like from a Dart
developer point of view.
For the record, Angular 2 is looking great.

Post by Matthew Butler
On Thursday, October 8, 2015 at 1:57:32 PM UTC-3, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.
Yes but the whole point of the discussion is for Dart 2.0. If it were to

remain 100% backward compatible it would be a minor version update, not a
major version number change.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
As of Java 6 (I'm not sure about later, I haven't looked), java

internal version was still 1.6.x (eg everything from Java 2 and up were
only minor version rolls, not a major version. This may have change since
that time)
--
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

--
Saudações,
Eduardo Teixeira Dias
------------------------------------------
Tendencies Consultoria Ltda.
Tel: [phone]
Cel: 11 9 9246-4192

[Jan Mostert]

Quote:
100% backward compatibility shoud be a goal.
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.

This is the reason I dumped Python and PHP and perl for Java even though
Java is much more verbose with all its ceremonies - stuff that worked on
Java5 is still working on Java6 is still working on Java7 is still working
on Java8 even though my IDE is now giving me suggestions saying I should be
using the new Java8 syntax, but it still compiles and still runs fine.

I'm all for shiny new syntax, but keep the old syntax around for a little
longer while concurrently adding new features and adding suggestions /
warnings in the output that the existing syntax is now deprecated and will
be removed in Dart3. If breaking changes are going into Dart2, what is
stopping breaking changes from going into Dart3 ?

I've done work for two enterprise clients recently where they would rather
go with something horrific like JSF and have all the developers foam with
anxiety due to having to learn JSF, but it's stable and has Oracle
approval, so they are going with JSF rather than going with something that
might change in the next couple of years even though there might be tools
to migrate it or what have you.

Hogging the old syntax along from V1 to V2 will do a lot for Dart's
credibility in the Enterprise even though the hard-core Dart Gurus doesn't
like to hear that.
I personally have two projects using Dart, one just shy of 50k LOC and the
other one way over 150k LOC, it would make me sleep easier at night as well
knowing that I could upgrade to Dart2 and only get warnings for using old
syntax and having sufficient time to migrate stuff before Dart3 comes out
rather than an All Or Nothing approach.

Post by Gen
Do you fear any change in particular ?
I see potential for incompatible changes only with regard to the type
- "var" will stand for static type interference.
- "strong mode" is the new standard. This might break much in your code.
It remains to be seen how much breaks in the standard library and for other
people.
How much code that is rejected in strong mode is also bad code anyway ?
I agree with you that losing existing code is no option.
But I bet that the Dart team will not confront you with that case unless
there is a compelling reason which I do not see.

Post by Eduardo Teixeira Dias
If you go incremental you do not lose packages.
The programers time are used to do apps and not to write the basics all
over again...
The cost of fragmentation is high.
You want some package that only exists for Dart 1 but you also need a
package that only exists in Dart 2.
As you see the benefits of the incompatibility is enourmous...
It is important to look also if it is not possible, in some way, get the
benefit of the change in a compatible incremental way.
If you just say "We are considering an incompatible Dart 2" all managers
that I know of will wait to see what monster will that Dart 2 look like
before even considering investing in Dart 1.
If in the other hand you say "We are considering an amazing new Dart 2
that is backward compatiple". You know, is different. I can invest in Dart
1 now.
You can only look at AngularDart forum. The day the Angular announced the
future incompatible Angular2, no upgrade path. The conversations just
stoped. Why??? Let's see what this angular2 will look like from a Dart
developer point of view.
For the record, Angular 2 is looking great.

Post by Matthew Butler
On Thursday, October 8, 2015 at 1:57:32 PM UTC-3, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.
Yes but the whole point of the discussion is for Dart 2.0. If it were

to remain 100% backward compatible it would be a minor version update, not
a major version number change.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
As of Java 6 (I'm not sure about later, I haven't looked), java

internal version was still 1.6.x (eg everything from Java 2 and up were
only minor version rolls, not a major version. This may have change since
that time)

--

Post by Matthew Butler
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

--
Saudações,
Eduardo Teixeira Dias
------------------------------------------
Tendencies Consultoria Ltda.
Tel: [phone]
Cel: 11 9 9246-4192

--
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

[Eduardo Teixeira Dias]

I have 3 worries:

1- Package incompatibility (technical)
- Fragmentation on an already thin package base
- Not been able to use Dart1 packages in a Dart2 app

2- Customers codebases
- Instead of write new stuff (Investment), the need to rewrite code that
already works (Cost)

3- Perception (Comercial). I want to sell Dart as soon as possible.
- I need to be able to talk to a customer an say with a straight face:
- You can invest in Dart now.
- Your investment will be preserved
- Your team will be more productive and you will get the job done
investing less.
- The thing will be maintainable with low cost
- Is easy to get your team trained in Dart
- The same language can be used to:
- SPA / Rich Client
- Mobile Hybrid
- Server
- Mobile Native (In the Future)

Tools, documentation, tutorials are the key to Dart Success.

Do not put on the account of the developers the low adoption of Dart.

It is still hard to get things done in Dart.

It is not the sintaxe that are preventing Dart of achieving the scape
velocity.

It is already Better than Java and mixed with some good Perl parts (How
not to like it the way it is?)

Please, dear lord, give me a sign so I can believe... ...That Dart can be
used and sold for internal projects now!

...

Post by Gen
Do you fear any change in particular ?
I see potential for incompatible changes only with regard to the type
- "var" will stand for static type interference.
- "strong mode" is the new standard. This might break much in your code.
It remains to be seen how much breaks in the standard library and for other
people.
How much code that is rejected in strong mode is also bad code anyway ?
I agree with you that losing existing code is no option.
But I bet that the Dart team will not confront you with that case unless
there is a compelling reason which I do not see.

Post by Eduardo Teixeira Dias
If you go incremental you do not lose packages.
The programers time are used to do apps and not to write the basics all
over again...
The cost of fragmentation is high.
You want some package that only exists for Dart 1 but you also need a
package that only exists in Dart 2.
As you see the benefits of the incompatibility is enourmous...
It is important to look also if it is not possible, in some way, get the
benefit of the change in a compatible incremental way.
If you just say "We are considering an incompatible Dart 2" all managers
that I know of will wait to see what monster will that Dart 2 look like
before even considering investing in Dart 1.
If in the other hand you say "We are considering an amazing new Dart 2
that is backward compatiple". You know, is different. I can invest in Dart
1 now.
You can only look at AngularDart forum. The day the Angular announced the
future incompatible Angular2, no upgrade path. The conversations just
stoped. Why??? Let's see what this angular2 will look like from a Dart
developer point of view.
For the record, Angular 2 is looking great.

Post by Matthew Butler
On Thursday, October 8, 2015 at 1:57:32 PM UTC-3, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.
Yes but the whole point of the discussion is for Dart 2.0. If it were

to remain 100% backward compatible it would be a minor version update, not
a major version number change.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
As of Java 6 (I'm not sure about later, I haven't looked), java

internal version was still 1.6.x (eg everything from Java 2 and up were
only minor version rolls, not a major version. This may have change since
that time)
--
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

--

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

[Don Olmstead]

If there are big breaking changes in 2.0 there's nothing saying you can't
stick with 1.x. There could also be transformers to migrate the 1.x code
over to 2.x. If you're just relying on Google packages those should for
sure get updated.

[Joel Trottier-Hébert]

I just want to mention, at the last Dart Summit, Google said they wouldn't
put breaking changes without making scripts/tutorials to make the upgrade
easy. So backward compatibility shouldn't be too bad I think.

Post by Don Olmstead
If there are big breaking changes in 2.0 there's nothing saying you can't
stick with 1.x. There could also be transformers to migrate the 1.x code
over to 2.x. If you're just relying on Google packages those should for
sure get updated.
--
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

[Matthew Butler]

Post by Don Olmstead
If there are big breaking changes in 2.0 there's nothing saying you can't
stick with 1.x. There could also be transformers to migrate the 1.x code
over to 2.x. If you're just relying on Google packages those should for
sure get updated.

Even more than transformers, Bob specifically mentioned in the "Types on

Post by Don Olmstead
"We would absolutely provide a migration tool".

So then you're only looking at libraries that haven't been maintained as
not working. And they're subject to break unexpectedly anyways. Breaking
changes aren't bad at this time.

Matt

[Günter Zöchbauer]

On Thursday, October 8, 2015 at 6:57:32 PM UTC+2, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.

I also think that this is very important but if something is considered a
bad decision in retrospect, it should be fixed now instead of working
around for the next decades. The focus for Dart changed quite a bit
recently and carrying the legacy can become a big burden.
After version 2.0 breaking changes are to be avoided at all cost.
If the direction again needs to change notably after 2.0 then it's
probably time for something entirely new anyway.
There are strategies to soften the impact of breaking changes like
supporting old and new syntax simultaneously for a while and/or automatic
code conversion and probably others.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
For me, being able to to write modern Rich Client Web Apps, Mobile Apps
and Server APIs (Redstone) using the same language is amazing but backward
compatibility is essential to convince customers.
Customers like that.
Performance gains and new functionalities are always wellcome!
For companies with large Python 2 code bases it is problem not necessarily
for the developers...

Post by Gen
I have no Dart code to lose or to port.
So "100% backward compatibility" is certainly not on my list.

So, you wrote no Dart code worth porting, cool...
The standard library is already statically and strongly typed because that

Post by Gen
is the main reason for using Dart and not Javascript.
Only the compiler must change to support Dart 2.
With Dart 2, the compiler and build tools could use a language version
property declared somewhere. If it is absent then treat the code as Dart 1
code.
I do not remember having read something against Python 3 and the need for
breaking changes.
The link you gave mentions that only 13 % think that Python 3 was a
mistake and maybe some of these people think that Python 3 did not change
enough.
AFAIK, the Python 2to3 tool works well.
Besides, I wonder why List<int> and not ByteBuffer is used in dart:io.
Maybe it does not matter for performance, does it ?

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn
- JBoss Seam
- Incompatibility: Seam 2 / Seam 3
- AngularDart
- Obsolete from the start
- Python
- Incompatibility
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/
<http://www.google.com/url?q=http%3A%2F%2Fwww.randalolson.com%2F2015%2F01%2F30%2Fpython-usage-survey-2014%2F&sa=D&sntz=1&usg=AFQjCNGGpdDg-MVf0jn374_zuzOKyZXNoA>
Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need to
set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503
777 0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--

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

[Hoo Luu]

圚 2015幎10月9日星期五 UTC+8䞊午3:49:16GÃŒnter Zöchbauer写道

Post by Günter Zöchbauer
On Thursday, October 8, 2015 at 6:57:32 PM UTC+2, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.

I also think that this is very important but if something is considered a
bad decision in retrospect, it should be fixed now instead of working
around for the next decades.

+1.Generics is one of the huge pain points of java.For me,java,the language
that I have to use at work everyday, is popular,successful but unpleasant
to write.Dart can rescue me from java so i wish dart take warning from java
and become both a popular and pleasant-to-write language in the future.

Post by Günter Zöchbauer
The focus for Dart changed quite a bit recently and carrying the legacy
can become a big burden.

+1. Especially after giving up integrating VM into chrome,Dart should not
carry the legacy.

Post by Günter Zöchbauer
After version 2.0 breaking changes are to be avoided at all cost.
If the direction again needs to change notably after 2.0 then it's
probably time for something entirely new anyway.
There are strategies to soften the impact of breaking changes like
supporting old and new syntax simultaneously for a while and/or automatic
code conversion and probably others.

Agree. It's time to innovate and even do somewhat breaking changes.Do
radical changes after 2.0 will pay a high price (eg, py3K,perl6, PHP
7,etc).

Post by Günter Zöchbauer

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
For me, being able to to write modern Rich Client Web Apps, Mobile Apps
and Server APIs (Redstone) using the same language is amazing but backward
compatibility is essential to convince customers.
Customers like that.
Performance gains and new functionalities are always wellcome!
For companies with large Python 2 code bases it is problem not
necessarily for the developers...

Post by Gen
I have no Dart code to lose or to port.
So "100% backward compatibility" is certainly not on my list.

So, you wrote no Dart code worth porting, cool...
The standard library is already statically and strongly typed because

Post by Gen
that is the main reason for using Dart and not Javascript.
Only the compiler must change to support Dart 2.
With Dart 2, the compiler and build tools could use a language version
property declared somewhere. If it is absent then treat the code as Dart 1
code.
I do not remember having read something against Python 3 and the need
for breaking changes.
The link you gave mentions that only 13 % think that Python 3 was a
mistake and maybe some of these people think that Python 3 did not change
enough.
AFAIK, the Python 2to3 tool works well.
Besides, I wonder why List<int> and not ByteBuffer is used in dart:io.
Maybe it does not matter for performance, does it ?

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn
- JBoss Seam
- Incompatibility: Seam 2 / Seam 3
- AngularDart
- Obsolete from the start
- Python
- Incompatibility
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/
<http://www.google.com/url?q=http%3A%2F%2Fwww.randalolson.com%2F2015%2F01%2F30%2Fpython-usage-survey-2014%2F&sa=D&sntz=1&usg=AFQjCNGGpdDg-MVf0jn374_zuzOKyZXNoA>
Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need
to set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503
777 0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--

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

[kc]

Post by Günter Zöchbauer
On Thursday, October 8, 2015 at 6:57:32 PM UTC+2, Eduardo Teixeira Dias

Post by Eduardo Teixeira Dias
100% backward compatibility shoud be a goal.

I also think that this is very important but if something is considered a
bad decision in retrospect, it should be fixed now instead of working
around for the next decades. The focus for Dart changed quite a bit
recently and carrying the legacy can become a big burden.
After version 2.0 breaking changes are to be avoided at all cost.

Right I think the best approach for 2.0 is too lay the foundation (with
Flutter) for future enhancements which can be backward compatible manner. A
lot of 2.0 should about removing things.

K.

Post by Günter Zöchbauer
If the direction again needs to change notably after 2.0 then it's
probably time for something entirely new anyway.
There are strategies to soften the impact of breaking changes like
supporting old and new syntax simultaneously for a while and/or automatic
code conversion and probably others.

Post by Eduardo Teixeira Dias
It shows the entertprise that Dart is safe for investment.
I have Java code with more than 10 years that runs fine on current Java VMs.
For me, being able to to write modern Rich Client Web Apps, Mobile Apps
and Server APIs (Redstone) using the same language is amazing but backward
compatibility is essential to convince customers.
Customers like that.
Performance gains and new functionalities are always wellcome!
For companies with large Python 2 code bases it is problem not
necessarily for the developers...

Post by Gen
I have no Dart code to lose or to port.
So "100% backward compatibility" is certainly not on my list.

So, you wrote no Dart code worth porting, cool...
The standard library is already statically and strongly typed because

Post by Gen
that is the main reason for using Dart and not Javascript.
Only the compiler must change to support Dart 2.
With Dart 2, the compiler and build tools could use a language version
property declared somewhere. If it is absent then treat the code as Dart 1
code.
I do not remember having read something against Python 3 and the need
for breaking changes.
The link you gave mentions that only 13 % think that Python 3 was a
mistake and maybe some of these people think that Python 3 did not change
enough.
AFAIK, the Python 2to3 tool works well.
Besides, I wonder why List<int> and not ByteBuffer is used in dart:io.
Maybe it does not matter for performance, does it ?

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
2- Performance enhancements to compete with: Python, Java, PHP on the
server
3- Only break rule 1 if it is needed to boost performance
4- Be Pragmatic
- Perl 6
- JBoss Portal
- Incompatibility: JBoss Portal / GateIn
- JBoss Seam
- Incompatibility: Seam 2 / Seam 3
- AngularDart
- Obsolete from the start
- Python
- Incompatibility
http://www.randalolson.com/2015/01/30/python-usage-survey-2014/
<http://www.google.com/url?q=http%3A%2F%2Fwww.randalolson.com%2F2015%2F01%2F30%2Fpython-usage-survey-2014%2F&sa=D&sntz=1&usg=AFQjCNGGpdDg-MVf0jn374_zuzOKyZXNoA>
Python 3 was released 12/03/2008 yet more code is written nowadays in
Python 2. Library compatibility hell.

Post by kc
Lets hope Dart 2 isn't a Perl 6.
K.
On Wednesday, October 7, 2015 at 11:05:18 PM UTC+1, Randal L. Schwartz

Filipe> I would add that it becomes even more awesome when you need
to set multiple
Filipe> final element = new DivElement()
Filipe> ..id = 'errorId'
Filipe> ..className = 'errorStyle'
Filipe> ..text = 'Some Message'
Filipe> ..title = 'Some Title';
And awesomely, the Dart cascade got it right, which was always a bit of
a wart in Smalltalk, requiring many cascades to end in "; yourself."
--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503
777 0095
Perl/Unix consulting, Technical writing, Comedy, etc. etc.
Still trying to think of something clever for the fourth line of this .sig
--

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

--

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

[Yegor]

Post by Eduardo Teixeira Dias
1- 100% backward compatibility
- Perl 6
- JBoss Portal
- JBoss Seam
- AngularDart
- Python

Then let's learn from the best. golang (remember gofix?) and swift come to
mind first. Both caused minimal disruption and both are very successful
languages in their domains.

[Eduardo Teixeira Dias]

The development of Perl 6 started at 2005. and seems that the official
release will take place later this year... 10 years in the making.A large
part of active Perl5 users does not see Perl 6 as an upgrade path for Perl
5 projects.
For the time to market point of view and from the point of view of
community cohesion it is a failure. A new (possible even successfull
)language if you will...
I prefer the Java aproach. No dead bodies all over the way...
The zombies remember...

I agree with the sugestion of: Jan Mosterf
"Hogging the old syntax along from V1 to V2 will do a lot for Dart's
credibility in the Enterprise even though the hard-core Dart Gurus doesn't
like to hear that.I personally have two projects using Dart, one just shy
of 50k LOC and the other one way over 150k LOC, it would make me sleep
easier at night as well knowing that I could upgrade to Dart2 and only get
warnings for using old syntax and having sufficient time to migrate stuff
before Dart3 comes out rather than an All Or Nothing approach."In this way
we can immediatly upgrade our projects do Dart2 as soon as it came out and
then you prevent fragmentation.
As an example: If I have a CRM application and a need a new module. Dart2
comes out. What shoud I do? Start developping a new codebase in Dart2
Server and a new codebase in Dart2 Angular.
- On the server a new process in another port a new mapping on Apache -
On the client smach the new component some way into the early componentsOr
we take a hit to upgrade the entire codebase, test everything and then
start the new development (Adicional Cost in time and money).
The linux aproach is good too. Just incremental changes sinse, I can't even
remember, and sometimes Linus just upgrade the major number for no reason
at all.
"It seems very likely that the recent 2.6.39 Linux kernel was the last
major release of the 2.6.x series.""Linus Torvalds made a remark on the
Linux Kernel Mailing list at the beginning of the week that the voices were
telling him that the numbers were getting too big... and they are..aren't
they?"
"More than the numbers inside of 2.6, the time has come for change, for a
number of reasons. Torvalds gives at least one good reason related to
timing:"
In some cases Boring is good.

[kc]

Ideally a Dart 2.0 could be released with Flutter. Remove boilerplate and
ceremony and dial-up the expressiveness (within the constraint of the
Flutter schedule). A kind of relaunch.

K.

Any info available on how an Android app coded in Dart/Sky compares to
'native' Java/XML in terms of performance and resource usage
(mem/cpu/battery) etc.
I bang on about syntax/semantics but if performance is ballpark to
'native' then Dart could be very successful regardless (and give a catalyst
to dart2js). The dev experience will be so much nicer. (Though fix the
Java-ism's pls).
Dart was outflanked by ES6/TS on the web but could succeed via Sky/mobile.
K.