Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!seismo!mcvax!enea!liuida!gorry
From: gorry@liuida.UUCP
Newsgroups: comp.lang.smalltalk
Subject: Re: Messages?
Message-ID: <577@tragicomix.liu.se>
Date: Sun, 7-Jun-87 09:30:53 EDT
Article-I.D.: tragicom.577
Posted: Sun Jun  7 09:30:53 1987
Date-Received: Sun, 7-Jun-87 23:39:44 EDT
References: <572@tragicomix.liu.se> <574@tragicomix.liu.se> <tragicom.576> <283@suadb.UUCP>
Organization: CIS Dept, Univ of Linkoping, Sweden
Lines: 20

In reply to my request
>Although the concept of inheritance is very powerful, it is clearly
>associated with overhead from running around hierarchies to find
>methods...

anders@suadb.UUCP (Anders Bj|rnerstedt) wrote:
>>Inheritance as such is not the cause of overhead. Dynamic binding
>>of methods is. 

Naturally, if there is some way of providing static binding through a
compiler, the problem I mentioned with inheritance certainly would not
arise, as the method lookup would be replaced by a subroutine call.

So, in the case of dynamic binding, inheritance sure adds to the
overhead of finding a method. A typical implementation looks up the
method from the hashtable belonging to the class of the object,
follows the inheritance chain if not found, and repeats. To shorten
these paths, some redundancy scheme might be applicable.

gorry						gorry@majestix.liu.se