I don't get it

Maybe Trygve will explain? Not sure if he's reading the newbies list though. CC'ing.

- Bert -

On 21.06.2011, at 16:27, Ron Teitelbaum wrote:

Hi Erlis,

 

Class based language is very difficult for some people to grasp.  I’ve trained a number of programmers.  There is this AH moment when people get it.  Before that the idea is just so foreign to them that they have a difficult time grasping what is going on.  I would often get the question, “but where do I put the code”.  The answer is everywhere.  Thinking in terms of objects and programming their behavior makes sense.  People get that.  Make a Robot and teach it how to walk.  If I tell the robot to walk she does.  Ok great.  Now I have all these objects that do stuff where do I put the code?  This is where people stumble.  In my opinion the leap to understanding is very short.  The answer is simple.  (Answer: It depends on how you want to interact with the user). 

 

I do not believe that anything is screwed up.  It is just a conceptual model that requires AH!

 

All the best,

 

Ron Teitelbaum

Immersive Collaboration Expert

3d Immersive Collaboration Consulting

[hidden email]

Follow me on Twitter

twitter.com/RonTeitelbaum

www.3dicc.com

 

 

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Erlis Vidal
Sent: Tuesday, June 21, 2011 9:47 AM
To: smalltalk
Subject: [Newbies] I don't get it

 

Hi guys, 

I was reading recently Jim Coplien ideas about DCI and I was surprised with the following statement: 

"The Smalltalk people, when they put together Smalltalk originally the computational model was exactly right, in terms of thinking in terms of objects. And then they screwed it up with the language. The language is so class focused! Most languages that we’re saddled with today had made this error."

you can find it here: http://blog.redtexture.net/2010/06/01/coplien-on-dci-mvc/

Does someone knows what Jim is criticizing? In which sense the computational model was screwed by the language? I see smalltalk as a language that express very well the intended computational model behind. 

But that's just me, maybe some of you could help me to understand Jim's point. 

Thanks
Erlis

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

----- Original Nachricht ----
Von:     Erlis Vidal <[hidden email]>
An:      smalltalk <[hidden email]>
Datum:   <a href="tel:21.06.2011%2015" value="+12106201115">21.06.2011 15:46
Betreff: [Newbies] I don't get it

> Hi guys,
>
> I was reading recently Jim Coplien ideas about DCI and I was surprised with
> the following statement:
>
> *"The Smalltalk people, when they put together Smalltalk originally the
> computational model was exactly right, in terms of thinking in terms of
> objects. And then they screwed it up with the language. The language is so
> class focused! Most languages that we?re saddled with today had made this
> error."
> *
> you can find it here:
> http://blog.redtexture.net/2010/06/01/coplien-on-dci-mvc/
>
> Does someone knows what Jim is criticizing? In which sense the
> computational
> model was screwed by the language? I see smalltalk as a language that
> express very well the intended computational model behind.
>
> But that's just me, maybe some of you could help me to understand Jim's
> point.
>
> Thanks
> Erlis
>
>

> --------------------------------

>
> _______________________________________________
> Beginners mailing list
> [hidden email]
> http://lists.squeakfoundation.org/mailman/listinfo/beginners
>

You find the answer on the page the link to you cited above:

"... It is interesting that (as a consequence of web browsers) the most used language on the planet, JavaScript, satisfies his desire for a non-class-based object-oriented programming language.
..."

To explain it in short: There are two paradigms of Object orientation, a class-based one and a prototype based one. The cleanest class-based one is smalltalk. There are some protype-based languages, with javascript as the most popular one - but most people do not even know that JavaScript is Prototype-based. A prototype-based language, does not make the distinction of classes vs objects: it simply has objects.

"A prototype-based language has the notion of a prototypical object, an object used as a template from which to get the initial properties for a new object. Any object can specify its own properties, either when you create it or at run time. In addition, any object can be associated as the prototype for another object, allowing the second object to share the first object's properties." (from  http://stackoverflow.com/questions/186244/what-does-it-mean-that-javascript-is-a-prototype-based-language)

There are some more prototype based languages, as for instance "io", derived from self. A new and interesting prototype one is Lisaac (http://en.wikipedia.org/wiki/Lisaac). If you are interested in an overview of prototype-based languaged see http://en.wikipedia.org/wiki/Prototype-based_programming.

Let me first add, that the "siver bullet" of programming has not yet been found. From a system design point of view it has been recognized, that class based design leads to some "rigidification" in growing systems. If you study the pototyping way of OOP you will find a much higher plasticiy of the system. Obviously that is what James Coplien is missing in class-based programming. (I am quite sure that many Smalltalk users will disagree. But do Smalltalk programmers really know the prototype part of JavaScript (not just the DOM) in depths? Or other prototype based languages?  As a rule  that is not the rule.)

To my regret I have to add, that most of the CS-professors do not know the difference. I had a lot of trouble with this topic in the past. May be things are different in OO-research active countries.

Any flames?

Have fun!

B. Blochl

PS: A personal unsolicited advice of an old man: Do not lose any time with procedural languages! (Possible exception: You have to program close to the processors hadware.  because , to my regret, register are still now not objects. )

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

On 22.06.2011, at 12:05, David Corking wrote:

> Erlis  wrote:
>
>> Now I think I start understanding Jim statement, basically the statement is:
>> Smalltalk (the language) adopted the classical way of OO instead the
>> prototype way, so I don't have all the plasticity I was expecting...
>>
>> Do the smalltalkers feel this way?
>
> Perhaps. But Squeakers have access to other languages: notably Etoys.
> It isn't often mentioned that Etoys is an implementation of the Self
> computation model as a visual programming language for children.
> Interestingly, almost any Morphic class can easily become an Etoys
> object. (This won't be surprising to those who know that Morphic was
> first built for Self.)

Indeed, and thanks for pointing it out. Adding instance-specific methods isn't hard to do in Squeak (Etoys does it), although modifying method-lookup to implement delegation is more challenging.

It seems, however, that prototypes is not really what Jim and Trygve are aiming at. The paper instead mentions Traits, which indeed add a lot of flexibility to the traditional Smalltalk class model:

       http://www.artima.com/articles/dci_vision.html

- Bert -

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

Am <a href="tel:21.06.2011%2015" value="+12106201115" target="_blank">21.06.2011 15:46, schrieb Erlis Vidal:

Hi guys,

I was reading recently Jim Coplien ideas about DCI and I was surprised with the following statement:

"The Smalltalk people, when they put together Smalltalk originally the computational model was exactly right, in terms of thinking in terms of objects. And then they screwed it up with the language. The language is so class focused! Most languages that we’re saddled with today had made this error."

you can find it here: http://blog.redtexture.net/2010/06/01/coplien-on-dci-mvc/

Does someone knows what Jim is criticizing? In which sense the computational model was screwed by the language? I see smalltalk as a language that express very well the intended computational model behind.

But that's just me, maybe some of you could help me to understand Jim's point.

Thanks
Erlis

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

As I actually found, there already is an implementation of a DCI-System in Squeak/Smalltalk. You can find an article, a documenation and a download on
http://heim.ifi.uio.no/~trygver/themes/babyide/babyide-index.html

Also given is a DCI-Maillist:
object-composition (at) googlegroups.com.

One might be interested in "The Common Sense of Object Oriented Programming" by Trygve Reenskaug
http://heim.ifi.uio.no/~trygver/2009/commonsense.pdf

Regards

B. Blochl

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners

Erlis  wrote:

> Now I think I start understanding Jim statement, basically the statement is:
> Smalltalk (the language) adopted the classical way of OO instead the
> prototype way, so I don't have all the plasticity I was expecting...
>
> Do the smalltalkers feel this way?

Perhaps. But Squeakers have access to other languages: notably Etoys.
It isn't often mentioned that Etoys is an implementation of the Self
computation model as a visual programming language for children.
Interestingly, almost any Morphic class can easily become an Etoys
object. (This won't be surprising to those who know that Morphic was
first built for Self.)

B. Blochl wrote:

> But I think all that is a bit off topic on a Smalltalk mail list. But to remember: You had a special

> Croqueford -question I tried to clarify. I think the future mails [s]hould focus on SmallTalk topics.

Not really, in my opinion: this is a Squeak mailing list - not a
Smalltalk one. Squeak has been a fertile platform for language
experimentation. I would love to hear about other classless languages
that have been implemented on Squeak: I think Scratch might be one.

David

p.s. Off-topic: CLOS* is an interesting language inspired by
Smalltalk-80, which is so flexible you might want to think of it as
classless. I have never programmed in it, but its proponents sound
enthusiastic about its malleability.
* 'either one syllable rhyming with "dross", or two syllables as in
"see-loss" ... [W]ith the exception of slot accessors - all of your
application's functionality lives in function and method definitions,
not in classes. It's sometimes appropriate to place methods applicable
to some class into the same file as that class. It's sometimes
appropriate to place all the methods of a generic function into a
single file. There are no language constraints on this. ': Nick Levine
http://cl-cookbook.sourceforge.net/clos-tutorial/index.html

_______________________________________________
Beginners mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/beginners