Subj : Re: current state of affairs wrt __proto__
To   : netscape.public.mozilla.jseng
From : Harald Albrecht <Harald.Albrecht@plt.rwth-aachen.de>
Date : Sun Jan 19 2003 05:54 pm

Waldemar Horwat wrote:
>> Is there a reason why the ECMA spec does not contain this kind of 
>> information?

> I don't understand.  Please clarify what kind of information you're 
> asking about.

Waldemar,

please excuse me for not coming back to your legitimate request for my 
spec moaning.

I'm not sure what the intented audience of the ECMAscript specification 
is. For my moaning I assume both implementers as well as programmers, 
who want to get genuine information "straight from the horse's mouth".

In case of specifications I have come to appreciate it if a 
specification first starts with an overview from a more abstract point 
of view without all the details. For instance, in the case of ECMAscript 
specification, I would be pleased if the spec would first start with an 
overview on what the basic design ideas are, with some specific parts 
like prototype-oriented issues, highlighted. In my oppinion, your paper 
on ECMAscript 2.0 does exactly this: it describes important features of 
ECMAscript (in this case even for 1.x) without too much syntactic sugar 
and nitty-gritty details that would distract the reader from the core 
ideas. For instance, what is a prototype, how does "inheritance" work 
here, what features does an object have in any case, how does the 
prototype slot interact with a constructor function etc. Not in all the 
gory details, that is already laid down in the existing specification. 
But rather for a layman, so he can start to understand the design of 
ECMAscript and can start to use it the way it was meant to be.

To some extend, the rhino book also tries to go this way, but it already 
goes too much into browser-related things and has some problems with 
terminology (for instance, using class-oriented terminology for a 
prototype-based language). Also, the rhino book does not deal properly 
with the language and its design decissions but rather focuses on 
writing browser JavaScript code (understandably).

On a sidenote: I'm really interested to see when ECMAscript 2 
implementations will enter the browser scene as standard browser parts. 
ECMAscript 2 looks very promising and it will be interesting to see how 
the integration of class-based and prototype-based aspects works on the 
large scale.

-- Harald

.