Newsgroups: comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!nntp.club.cc.cmu.edu!hudson.lm.com!news.pop.psu.edu!psuvax1!uwm.edu!cs.utexas.edu!utnut!nott!cunews!freenet.carleton.ca!FreeNet.Carleton.CA!ah004
From: ah004@FreeNet.Carleton.CA (Gregg MacDonald)
Subject: Smalltalk (simple? question)
Message-ID: <D1yML4.8vx@freenet.carleton.ca>
Sender: ah004@freenet2.carleton.ca (Gregg MacDonald)
Reply-To: ah004@FreeNet.Carleton.CA (Gregg MacDonald)
Organization: The National Capital FreeNet
References: <knight.789350186@tina.mrco.carleton.ca> <3cl83v$jmq@news1.delphi.com> <1995Jan5.094541.16980@wavehh.hanse.de>
Date: Fri, 6 Jan 1995 01:24:40 GMT
Lines: 24


In a previous posting, Alan Knight (knight@mrco.carleton.ca) writes:
> In <1995Jan5.094541.16980@wavehh.hanse.de> cracauer@wavehh.hanse.de (Martin Cracauer) writes:
> 
> 
>>Using the right or a wrong language for a given task will cause such
>>differences every time, of course. The questions is, how much areas
>>are there where C++ or Smalltalk is the right language. Would you
>>implement your Smalltalk compiler using Visualworks or xlisp?
> 
> I'm not quite sure what your point is here, but all of the Smalltalk
> compilers I know of are implemented in entirely in Smalltalk.
> 

A semi-related question: I know that the underlying layer of Smalltalk is
a Virtual Machine of some sort... does this apply to typical IBM
implementations (DOS/Win16/Win32/NT/OS2) or is the bottom-most layer in
the native machine language (386 I assume)... If it is implemented as a
VM, wouldn't there be much greater performance if it was done natively
(esp. primitives)?

ttyl,
Gregg

