Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!rice!uw-beaver!uw-june!ewan
From: ewan@cs.washington.edu (Ewan Tempero)
Newsgroups: comp.object
Subject: Re: object-oriented this, that, and the other thing
Message-ID: <10028@june.cs.washington.edu>
Date: 1 Dec 89 21:12:04 GMT
References: <2426@umbc3.UMBC.EDU> <190@ark1.nswc.navy.mil> <1561@novavax.UUCP> <21986@brunix.UUCP>
Organization: U of Washington, Computer Science, Seattle
Lines: 55

-Ted
>me

In article <21986@brunix.UUCP>, twl@brunix (Ted "Theodore" (W) Leung) writes:
-In article <10013@june.cs.washington.edu> ewan@june.cs.washington.edu 
-(Ewan Tempero) writes:
->	(a) Ways to create objects
->	(b) Ways to manage objects
->	(c) no non-object constucts [ill-defined I know, maybe I'll 
->					say something later]

-So what makes a language that supports these features any different
-from a language that is based purely on abstract data types?

This is a good question but I'm going to wimp out and not answer it
directly (for now, anyway). My "definition" was only intended for a
discussion on whether "classes" were required for OO languages, that
is, it was to be taken in the context of OO languages only (afterall,
I didn't define what an object was so that's the only context that
makes sense...if sense is to be made from what I wrote). Which brings
me to:-
-In the
-absence of code-sharing/grouping mechanisms, what distinguishes an
-object from an ADT?

The point I was trying to make, was that there is nothing inherent
in objects that relates to code-sharing mechanisms. I read somewhere
that it would be possible to put environment support on top of
Modula-2 (or was it Ada?) that would allow things of the form
NewModule.NewFunction = OldModule.OldFunction + stuff [this is very
hazy and I don't have a reference for it. Does anyone?]. This
is essentially what Simula does. As I have already said, there are
languages that I think anyone would say provide objects and ways
to manipulate them, but don't have code-sharing. They are also based
purely on ADTs (so OO implies ADTs, at least). If you are going to
insist that code-sharing mechanisms must be present then I think
"object" is a misnomer.


-This opens up the age old question of "what is an
-object", which will probably cause this group to go into conniptions.

:-) But if we can't discuss that here, where do we? :-) What I'm trying
to do, is to answer exactly this question, without the baggage of
code sharing (or concurrency, or anythings else). I think the concept
of "object" exists independently and, once we have a better understanding
of what this concept is, we can worry about the other stuff.

-I don't think that labelling all the features with the right
-object-oriented whatever is what's needed.  I think that we need to
-get good definitions of what the essential language elements are and
-try to understand their interactions in the language design space.

...ahem. Yes, I agree. You said it better...
--ewan


Brought to you by Super Global Mega Corp .com