Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!munnari!mimir!wacsvax!swanee!gustav
From: gustav@swanee.OZ (Gustav)
Newsgroups: comp.sys.mac
Subject: MPW and MultiFinder
Message-ID: <359@swanee.OZ>
Date: Fri, 4-Sep-87 22:57:21 EDT
Article-I.D.: swanee.359
Posted: Fri Sep  4 22:57:21 1987
Date-Received: Fri, 11-Sep-87 00:56:25 EDT
Organization: El. & El. Eng., Uni. of Western Australia
Lines: 37
Keywords: MPW, background jobs

In article <3344@apple.UUCP> lsr@apple.UUCP (Larry Rosenstein) writes:

> MPW 2.0 is Multifinder aware; not only does it work with Multifinder, but it
> takes vantage of the Multifinder environment.  
> For example, if you are running Multifinder and launch an application from
> MPW, the MPW shell does not run its normal suspend script.  It simply uses
> sublaunching to have Multifinder run the application in a separate area of
> memory (exactly as if you had launched the app from the Finder).  Since MPW
> does not exit, there is no reason to save the state of variables, aliases,
> etc. [...] [but] you can't run the compilers in the background under 
> Multifinder.  

Well, that's not good enough. Multitasking is of greatest value
exactly in those areas where you don't do things interactively and
compilations are some of them. Lengthy mathematics (it's quite commmon
that some Monte Carlo simmulations would run for DAYS!) is another
area. And basically MPW should be able to do all that. 

In fact one really wonders, Apple at this stage seems to support three
different operating environments, which does lead to some questions.
You have MultiFinder (or Finder in its older version), MPW, which
really stands on its own and A/UX. MPW has some features in common
with UNIX, and probably MPW shell or its derivative could be
implemented under UNIX, but basically it's a different thing. Now, to
support three different environments is a rather costly exercise, not
to mention a great potential for confusion amongst programmers and
users. Is Apple really going to do that? A word of advice about the
future plans in this respect would certainly help. 

In my opinion the best solution would be to amalgamate somehow all
three of them, which would perhaps provide a sort of an "ideal" OS for
a microcomputer. On the other hand, for a common user, which does not
program him/her-self, that might be an overkill (and over-cost). Also,
such an amalgamated system may require much more memory and disk space
than many users at this stage would be prepared to pay for.

Hmm... Have I just answered my own question?