From: Ilias Lazaridis <ilias@...>
Date: 2005-04-19T12:34:34+09:00
Subject: Re: - E03c - The Ruby Object Model (Revised Documentation)

Lionel Thiry wrote:
> Ilias Lazaridis a �crit :
>> Nikolai Weibull wrote:
>>> Ilias Lazaridis, April 18:
>>>> One should call Mr. Matz to review this.
>>>
>>>> Most possibly he's the only one who knows.
>>>
>>> It seems that you don't anyway,
>>
>> I've invested some time (whilst ignoring finally the current 
>> documentation) to create this UML diagramm:
>>
>> http://lazaridis.com/case/lang/ruby/TheRubyObjectModel.png
>>
>> I'm asking now for confirmation/feedback.
>>
>> This is nothing special within serious software-development.
>>
>> -
>>
>> But as it looks, this community is incapable to do so (giving 
>> productive feedback).
> 
> I can.
> 
> -
> 
> But there are other ways of asking, you know... :(

there are other ways of answering, too.

> -
> 
> I'll give it a try: the documentation "ri Class" gives is not wrong at 
> all. Maybe not complete, but not wrong.

it's incomplete, wrong and missleading.

> -
> 
> In "ri Class" text, if you replace references to "metaclasses" (which 
> are said to be between parentheses) by "singleton method placeholders", 
> perhaps it becomes clearer?

no, then it becomes completely confusing.

> 
> -
> 
> I remember that what was disturbing you was the stuff about metaclasses. 
> I told you ruby invite you to think more in terms of singleton method 
> than in terms of metaclasses. But it seems that you ignore what are 
> singleton methods.

of course i ignore them.

they are not relevant when reviewing ruby's oject-model.

> 
> -
> 
> Have you ever evaluated some classless OO langages? I know some: IO ( 
> http://www.iolanguage.com/ ), Self ( http://c2.com/cgi/wiki?SelfLanguage 
> ) and REBOL (rebol is not basically OO, but its OO part is classless. 
> http://www.rebol.com/ ). If you would, the concept of singleton method 
> would be far easier to understand for you. (this is actually how I 
> understood that concept)

[will possibly take a look at them shortly]

> 
> -
> 
> http://lazaridis.com/case/lang/ruby/TheRubyObjectModel.png
> 
> Your UML diagram of ruby object model is hard to read, and probably 
> wrong. The separation between classes in memory and in source code is 
> totally confusing. 

the seperation exist in reality (and in the "ri class" documentation).

additionally: It was neccessary for my understanding process.

the next version (1.1), the diagramm will be simplified (possibly in a 
few hours, if the stuff has passed).

> In ruby, classes *are* objects. 

yes, but there is still a class definition.

> Learn more about 
> classless languages, and it should be far easier to understand these 
> concepts.

this is not relevant to understand ruby's Object Model.

I just need a (correct) UML diagram to understand Ruby's Object Model.

> 
> There is another evident mistake in your UML diagram. Object doesn't 
> inherit from nil, nil is a literal and an instance of NilClass which 
> inherit from Object.

you are wrong:

irb(main):001:0> Object.superclass
=> nil
irb(main):002:0> nil.class
=> NilClass

[btw: to simplify the analysis, I've stopped exporing "nil" at that point]

> 
> -
> 
> I tried my best, I can't say anymore. Please, don't ask me for lengthy 
> exhausting explanations. Try the languages I told you about, they are 
> really worth it.

will possibly do this shortly.

> 
> -
> 
> Have a nice day.
> 
> -- 
> Lionel Thiry
> 
> PS: Please, anybody, do not blame me for anwsering this post. Do not 
> blame me for imitating Lazaridis posting. Do not blame me for trying to 
> be nice.

..

-- 
http://lazaridis.com