Re: The Foundation of OO (XDb)

From: Christopher Browne <cbbrowne_at_acm.org>
Date: 14 Jun 2002 05:40:35 GMT
Message-ID: <aebvki$5pd2h$1_at_ID-125932.news.dfncis.de>


Oops! jraustin1_at_hotmail.com (James) was seen spray-painting on a wall:

>>   "The Self language doesn't have any classes. When we want a new
>>    object, we find an existing one and copy it. We can then change the
>>    copy, safe in the knowledge that we have not affected anything
>>    else."

>> But note that this means that Self _rejects_ the notion of
>> "classes." It being an object-oriented system, that means that >> "classes" cannot be an intrinsic thing to OO.

> For the freedom of making isolated copies of an object, the loss of
> property and method inheritance seems costly in the general case,
> but may be advantages in certain circumstances, i think.

I am not making any inferences about the cost of anything, or of the advantages or disadvantages of anything. I couldn't care less about that; it's completely irrelevant.

The point I'm making is that some OO languages _require_ that objects be associated with classes, and that other OO languages have no classes to associate objects with.

Which means that making general claims about how classes are or are not associated with objects is an illegitimate thing to do.

-- 
(reverse (concatenate 'string "ac.notelrac.teneerf_at_" "454aa"))
http://cbbrowne.com/info/emacs.html
Rules of the Evil Overlord #122.  "The gun turrets on my fortress will
not  rotate enough  so that  they may  direct fire  inward or  at each
other. <http://www.eviloverlord.com/>
Received on Fri Jun 14 2002 - 07:40:35 CEST

Original text of this message