Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!nrl-cmf!ukma!husc6!purdue!decwrl!csdpie.dec.com!vilot
From: vilot@csdpie.dec.com (Michael J. Vilot)
Newsgroups: comp.lang.c++
Subject: RE: object data type
Message-ID: <8812301705.AA14595@decwrl.dec.com>
Date: 30 Dec 88 17:05:28 GMT
Organization: Digital Equipment Corporation
Lines: 36


 
> I am curious if anyone has a problem with the fact that c++ has a mixture of
> object data types, and standard C data types. I have just been looking at
> smaltalk and its philosophy of 'everything is an object'.

This turns out to be an interesting question.  
What you'll find in SmallTalk _implementations_ are special optimisations
for "primitive" data types (e.g. integers) to make them efficient enough for
traditional hardware.  C++ inherits C's "systems programming" orientation, and
naturally implements these "primitive" types efficiently:

	"Features that would incur run-time or memory overheads even when
	not used were avoided.  For example, ideas that would make it 
	necessary to store ``housekeeping information'' in every object were
	rejected;  if a user declares a structure consisting of two 16-bit
	quantities, that structure will fit into a 32-bit register."
	-- The C++ Programming Language, p. 3

>  [ question on generic queue ] 
> how does one work on datatypes such as integers,
> or doubles etc. Does one have to create a function that converts all defined
> datatypes to objects?		             

Not _all_.  Implicit and explicit conversions can simplify matters.
The arithmetic conversions of R6.6 (p. 253) are defined to cope with the
built-in types.  You can provide explicit conversions [R7.2.3, p. 258] in any
user-define classes.

Take the trouble to learn the conversion rules.  Like the operator precedence
hierarchy, a clear understanding of them can often simplify your code.
==
Mike Vilot			
e-mail:	       			vilot%ontologic@erl.mit.edu
or:				vilot%keeper.dec@decwrl.dec.com
when that fails:    		(508)467-6166 [for now]