Re: [polymer-dev] Re: My Naive Thoughts On Polymer 2, And The Google I/O 2017 Presentations

[Nick Pavlica]

Hi Karl,
  The first rule of tech, is that we will never agree on anything :) 

On Friday, May 26, 2017 at 4:34:09 PM UTC-6, Karl Tiedt wrote:
>
> How best would you develop a collection of extendable components while 
> avoiding classes?
>

  Honestly, I don't believe that anyone can answerer this, they can only 
work within the bounds of what they know, or whats forced upon them.
 

> Even 10 years ago libraries rolled their own class systems... 
>

   They were/I was doing it allot longer than 10 years ago... But I get 
your point, it's a style that works for many, and shouldn't be ignored.  
However, I don't believe that we can should ignore other approaches, and 
should try to accommodate them when and where it's possible.
 

> Should we avoid the platform way of doing inheritance and extensibility 
> just because?
>
Everything you have alluded to in the past is all making do with what was 
> available at the time... 
>

  No, we shouldn't ignore the platform, but we should also try to 
incorporate new ideas.  As I pointed out in the code above, there is little 
to no real innovation in Polymer and web components, with the exception 
that there are multiple vendors this time.  From the perspective of someone 
that's been doing this far too long, it just starts to feel like ground hog 
day.  

 

>
> ES6 is what's available from the platform...
>
With that being said, your aversion to it is confusing... 
>


  I don't know what's confusing about it.  ES6 is barley brining the 
language to a level of what was available decades ago, and it still has a 
Long way to go.  Thankfully things are starting to move forward, in ES, and 
other languages that give us some choice.

 

> I will agree that the "module" system that landed in ES6 is by far not the 
> smartest choice but it's what the lovely standards committee gave us...
>

  
  I think that Anders Hejlsberg did a great job in his presentation this 
year at MS Build, where he talks about how Typescript fits into the 
multi-paradigm hybrid thing that's Javascript/ES whatever.  He describes 
the use of classes in Typescript, but then reveals that the typescript 
compiler is actually written in a functional approach.  It's definitely 
worth the watch.  

  https://www.youtube.com/watch?v=d1f6VBmWg6o

If you haven't seen Rich Hickeys talk on Simplicity, it's a great watch, 
and may help clarify some of my thinking.

  https://www.youtube.com/watch?v=rI8tNMsozo0=11s
  
Regards!!
-- Nick

 

>
> On May 26, 2017 3:21 PM, "Nick Pavlica"  
> wrote:
>
>> Hi Bruce,
>>   Thanks for helping me better understand Polymer!
>>
>> Hi Bruce,
>>   Thanks for helping me better understand Polymer!
>>
>> On Friday, May 26, 2017 at 3:15:59 PM UTC-6, Bruce Anderson wrote:
>>>
>>> Hi Nick, I'm a fellow developer, not part of the Polymer team.  My first 
>>> foray into a production web app involved applets.  Glad we've moved on :-)
>>>
>>> But I don't think applets fell out of favor because they used classes.  
>>> That was the one aspect I've missed all these years, until recently.  They 
>>> were problematic in terms of performance, being confined to a rectangle, 
>>> blurry printing, needing to compile a complex, proprietary language when 
>>> JavaScript could do the same thing almost as fast, with a much lower 
>>> learning curve, etc.  And just not being "of the web" in the best sense, 
>>> but controlled by one vendor (who famously went after another vendor for 
>>> daring to ship their own version of it).
>>>
>>
>>   Classes, definitely weren't the cause of applets demise for sure :)   I 
>> was a huge Java/OOP zealot for years, and still have a copy of the "Object 
>> Oriented Thought Process" on my book shelf. Part of the point that I was 
>> trying to make, is that the OOP style isn't the only valid method of 
>> creating software, and we shouldn't always be pushing in that direction.  
>> The Polymer team/project are always talking about standards and the 
>> platform which makes me concerned that they aren't taking the other 
>> approaches into account.  This assumption is probably totally wrong, but 
>> it's the perception that I get from the docs, and the presentations that 
>> I've seen.  
>>
>>  
>>
>>>
>>> Most functional languages including F#, Scala, Haskell have classes.  
>>> Nothing is forcing you to make your classes have mutable members.  
>>>
>>
>>In a way, this helps clarify why I think that it's important to have, 
>> and encourage multiple styles.  There are pros and con's to each and we 
>> should be allowed to use the one we want or need to.
>>   
>>
>>>   
>>> I've seen examples of custom elements built with Elm 
>>> ,
>>>  
>>> for example.  You might also look at SkateJS, if you have a strong aversion 
>>> to OOP.
>>>
>>
>>   I've seem some of the Elm stuff, but haven't looked into SkateJS yet.  
>>
>>
>> Thanks again for the feedback!
>> -- Nick

Re: [polymer-dev] Re: My Naive Thoughts On Polymer 2, And The Google I/O 2017 Presentations

[Karl Tiedt]

How best would you develop a collection of extendable components while
avoiding classes? Even 10 years ago libraries rolled their own class
systems... Should we avoid the platform way of doing inheritance and
extensibility just because? Everything you have alluded to in the past is
all making do with what was available at the time...

ES6 is what's available from the platform... With that being said, your
aversion to it is confusing... I will agree that the "module" system that
landed in ES6 is by far not the smartest choice but it's what the lovely
standards committee gave us...

On May 26, 2017 3:21 PM, "Nick Pavlica"  wrote:

> Hi Bruce,
>   Thanks for helping me better understand Polymer!
>
> Hi Bruce,
>   Thanks for helping me better understand Polymer!
>
> On Friday, May 26, 2017 at 3:15:59 PM UTC-6, Bruce Anderson wrote:
>>
>> Hi Nick, I'm a fellow developer, not part of the Polymer team.  My first
>> foray into a production web app involved applets.  Glad we've moved on :-)
>>
>> But I don't think applets fell out of favor because they used classes.
>> That was the one aspect I've missed all these years, until recently.  They
>> were problematic in terms of performance, being confined to a rectangle,
>> blurry printing, needing to compile a complex, proprietary language when
>> JavaScript could do the same thing almost as fast, with a much lower
>> learning curve, etc.  And just not being "of the web" in the best sense,
>> but controlled by one vendor (who famously went after another vendor for
>> daring to ship their own version of it).
>>
>
>   Classes, definitely weren't the cause of applets demise for sure :)   I
> was a huge Java/OOP zealot for years, and still have a copy of the "Object
> Oriented Thought Process" on my book shelf. Part of the point that I was
> trying to make, is that the OOP style isn't the only valid method of
> creating software, and we shouldn't always be pushing in that direction.
> The Polymer team/project are always talking about standards and the
> platform which makes me concerned that they aren't taking the other
> approaches into account.  This assumption is probably totally wrong, but
> it's the perception that I get from the docs, and the presentations that
> I've seen.
>
>
>
>>
>> Most functional languages including F#, Scala, Haskell have classes.
>> Nothing is forcing you to make your classes have mutable members.
>>
>
>In a way, this helps clarify why I think that it's important to have,
> and encourage multiple styles.  There are pros and con's to each and we
> should be allowed to use the one we want or need to.
>
>
>>
>> I've seen examples of custom elements built with Elm
>> ,
>> for example.  You might also look at SkateJS, if you have a strong aversion
>> to OOP.
>>
>
>   I've seem some of the Elm stuff, but haven't looked into SkateJS yet.
>
>
> Thanks again for the feedback!
> -- Nick
>
>
>>
>>
>>
>> On Friday, May 26, 2017 at 4:08:20 PM UTC-4, Nick Pavlica wrote:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> *All,  After watching multiple presentations from Google I/O 2017 on
>>> Polymer 2, I was struck with an oppositional view of the “adopt the
>>> platform” mantra that the presenters were espousing.  While reflecting on
>>> this reaction for a couple of days, I couldn’t help but reminisce of days
>>> gone buy in this industry.  It feels as though we keep reinventing the same
>>> wheel, and expecting a different result.  For those of us that have been
>>> around long enough, we remember the failures of the past, and why they
>>> failed.  Specifically, I’m thinking of Java Applets, and all the time I
>>> wasted writing them.  While watching the presentations, I couldn’t help but
>>> notice the almost constant promotion of ES6 classes, and their use within
>>> Polymer 2.  I also couldn’t miss the constant assumption that this will be
>>> the standard, that should be adopted.  The more I thought about these
>>> annoyances, the more they continued to bother me, and it makes me question
>>> if our industry will every make any real progress, and move past a single
>>> set of ideas.  Here are some very basic examples to help highlight my
>>> point: A Basic Java Applet import java.applet.Applet;import
>>> java.awt.Graphics;public class HelloWorld extends Applet {public void
>>> paint(Graphics g) {g.drawString("Hello world!", 50, 25);
>>>}} A Simple Program Here is the
>>> output of my program:>> HEIGHT=25>  A Basic Polymer 2 Component >> rel="import"
>>>  href="https://polygit.org/components/polymer/polymer-element.html
>>> ">  //
>>> Define the class for a new element called custom-element  class
>>> CustomElement extends Polymer.Element {

[polymer-dev] Re: My Naive Thoughts On Polymer 2, And The Google I/O 2017 Presentations

[Nick Pavlica]

Hi Bruce,
  Thanks for helping me better understand Polymer!

Hi Bruce,
  Thanks for helping me better understand Polymer!

On Friday, May 26, 2017 at 3:15:59 PM UTC-6, Bruce Anderson wrote:
>
> Hi Nick, I'm a fellow developer, not part of the Polymer team.  My first 
> foray into a production web app involved applets.  Glad we've moved on :-)
>
> But I don't think applets fell out of favor because they used classes. 
>  That was the one aspect I've missed all these years, until recently.  They 
> were problematic in terms of performance, being confined to a rectangle, 
> blurry printing, needing to compile a complex, proprietary language when 
> JavaScript could do the same thing almost as fast, with a much lower 
> learning curve, etc.  And just not being "of the web" in the best sense, 
> but controlled by one vendor (who famously went after another vendor for 
> daring to ship their own version of it).
>

  Classes, definitely weren't the cause of applets demise for sure :)   I 
was a huge Java/OOP zealot for years, and still have a copy of the "Object 
Oriented Thought Process" on my book shelf. Part of the point that I was 
trying to make, is that the OOP style isn't the only valid method of 
creating software, and we shouldn't always be pushing in that direction.  
The Polymer team/project are always talking about standards and the 
platform which makes me concerned that they aren't taking the other 
approaches into account.  This assumption is probably totally wrong, but 
it's the perception that I get from the docs, and the presentations that 
I've seen.  

 

>
> Most functional languages including F#, Scala, Haskell have classes. 
>  Nothing is forcing you to make your classes have mutable members.  
>

   In a way, this helps clarify why I think that it's important to have, 
and encourage multiple styles.  There are pros and con's to each and we 
should be allowed to use the one we want or need to.
  

>   
> I've seen examples of custom elements built with Elm 
> ,
>  
> for example.  You might also look at SkateJS, if you have a strong aversion 
> to OOP.
>

  I've seem some of the Elm stuff, but haven't looked into SkateJS yet.  


Thanks again for the feedback!
-- Nick
 

>
>
>
> On Friday, May 26, 2017 at 4:08:20 PM UTC-4, Nick Pavlica wrote:
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> *All,  After watching multiple presentations from Google I/O 2017 on 
>> Polymer 2, I was struck with an oppositional view of the “adopt the 
>> platform” mantra that the presenters were espousing.  While reflecting on 
>> this reaction for a couple of days, I couldn’t help but reminisce of days 
>> gone buy in this industry.  It feels as though we keep reinventing the same 
>> wheel, and expecting a different result.  For those of us that have been 
>> around long enough, we remember the failures of the past, and why they 
>> failed.  Specifically, I’m thinking of Java Applets, and all the time I 
>> wasted writing them.  While watching the presentations, I couldn’t help but 
>> notice the almost constant promotion of ES6 classes, and their use within 
>> Polymer 2.  I also couldn’t miss the constant assumption that this will be 
>> the standard, that should be adopted.  The more I thought about these 
>> annoyances, the more they continued to bother me, and it makes me question 
>> if our industry will every make any real progress, and move past a single 
>> set of ideas.  Here are some very basic examples to help highlight my 
>> point: A Basic Java Applet import java.applet.Applet;import 
>> java.awt.Graphics;public class HelloWorld extends Applet {public void 
>> paint(Graphics g) {g.drawString("Hello world!", 50, 25); 
>>}} A Simple Program Here is the 
>> output of my program:> HEIGHT=25>  A Basic Polymer 2 Component > rel="import" 
>>  href="https://polygit.org/components/polymer/polymer-element.html 
>> ">  // 
>> Define the class for a new element called custom-element  class 
>> CustomElement extends Polymer.Element {static get is() { return 
>> "custom-element"; }constructor() {super(); 
>>this.textContent = "I'm a custom-element.";  }  }  // Register 
>> the new element with the browser  customElements.define(CustomElement.is, 
>> CustomElement);   > src="https://polygit.org/components/webcomponentsjs/webcomponents-loader.js 
>> <https://polygit.org/components/webcomponentsjs/webcomponents-loader.js>">
>>  
>> 
>>These feel eerily 
>> similar to each other don’t they?  I can’t help but think that a 

[polymer-dev] Re: My Naive Thoughts On Polymer 2, And The Google I/O 2017 Presentations

[Bruce Anderson]

Hi Nick, I'm a fellow developer, not part of the Polymer team.  My first 
foray into a production web app involved applets.  Glad we've moved on :-)

But I don't think applets fell out of favor because they used classes. 
 That was the one aspect I've missed all these years, until recently.  They 
were problematic in terms of performance, being confined to a rectangle, 
blurry printing, needing to compile a complex, proprietary language when 
JavaScript could do the same thing almost as fast, with a much lower 
learning curve, etc.  And just not being "of the web" in the best sense, 
but controlled by one vendor (who famously went after another vendor for 
daring to ship their own version of it).

Most functional languages including F#, Scala, Haskell have classes. 
 Nothing is forcing you to make your classes have mutable members.  

I've seen examples of custom elements built with Elm 
, 
for example.  You might also look at SkateJS, if you have a strong aversion 
to OOP.



On Friday, May 26, 2017 at 4:08:20 PM UTC-4, Nick Pavlica wrote:
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> *All,  After watching multiple presentations from Google I/O 2017 on 
> Polymer 2, I was struck with an oppositional view of the “adopt the 
> platform” mantra that the presenters were espousing.  While reflecting on 
> this reaction for a couple of days, I couldn’t help but reminisce of days 
> gone buy in this industry.  It feels as though we keep reinventing the same 
> wheel, and expecting a different result.  For those of us that have been 
> around long enough, we remember the failures of the past, and why they 
> failed.  Specifically, I’m thinking of Java Applets, and all the time I 
> wasted writing them.  While watching the presentations, I couldn’t help but 
> notice the almost constant promotion of ES6 classes, and their use within 
> Polymer 2.  I also couldn’t miss the constant assumption that this will be 
> the standard, that should be adopted.  The more I thought about these 
> annoyances, the more they continued to bother me, and it makes me question 
> if our industry will every make any real progress, and move past a single 
> set of ideas.  Here are some very basic examples to help highlight my 
> point: A Basic Java Applet import java.applet.Applet;import 
> java.awt.Graphics;public class HelloWorld extends Applet {public void 
> paint(Graphics g) {g.drawString("Hello world!", 50, 25); 
>}} A Simple Program Here is the 
> output of my program: HEIGHT=25>  A Basic Polymer 2 Component  rel="import" 
>  href="https://polygit.org/components/polymer/polymer-element.html 
> ">  // 
> Define the class for a new element called custom-element  class 
> CustomElement extends Polymer.Element {static get is() { return 
> "custom-element"; }constructor() {super(); 
>this.textContent = "I'm a custom-element.";  }  }  // Register 
> the new element with the browser  customElements.define(CustomElement.is, 
> CustomElement);src="https://polygit.org/components/webcomponentsjs/webcomponents-loader.js 
> <https://polygit.org/components/webcomponentsjs/webcomponents-loader.js>">
>  
> 
>These feel eerily 
> similar to each other don’t they?  I can’t help but think that a certain 
> style off code, primarily a certain form of OOPs is once again being forced 
> on the developer.  Each time the presenters proudly proclaimed we use ES6 
> classes with inheritance and everything;  I couldn’t help but cringe, then 
> follow it with a little chuckle.  I mean really, we were doing this two 
> decades ago.  With the ever increasing awareness of functional programming 
> styles and their benefits, it’s a little frustrating to feel like the 
> “platform” is forcing or promoting one style without an obvious open door 
> to other styles.  A monolithic conceptual paradigm will certainly result in 
> a false barrier to innovation, and will force the creation of even more 
> abstractions over the “platform”.   This just makes me want to stay as far 
> away from Polymer and web components as I can, which I know isn’t the 
> intention, but is the side effect of the message that was being presented. 
>  I’m certain that I’m not alone in the assumptions that I’ve made, and that 
> negative feelings that were inadvertently created as a result of these 
> presentations.   I’m sure that I’ve missed many of the finer details, and 
> I’m probably wrong on all accounts;  However,  I would love to learn how 
> the Polymer team, and the standards committees are planning to 
> interoperate, and even