Xref: utzoo comp.sys.mac:13910 comp.windows.misc:282
Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!pyramid!voder!apple!goldman
From: goldman@Apple.COM (Phil Goldman)
Newsgroups: comp.sys.mac,comp.windows.misc
Subject: Re: A/UX window systems, Mac toolbox, etc
Message-ID: <7670@apple.Apple.Com>
Date: 14 Mar 88 07:38:09 GMT
References: <4129@hoptoad.uucp> <283@rhesus.primate.wisc.edu> <1710@ssc-vax.UUCP> <9829@steinmetz.steinmetz.UUCP> <7593@apple.Apple.Com> <3609@bloom-beacon.MIT.EDU> <7656@apple.Apple.Com>
Reply-To: goldman@apple.UUCP (Phil Goldman)
Organization: Apple Computer Inc, Cupertino, CA
Lines: 74
Summary: Stop Picking on MultiFinder







In article <3690@bloom-beacon.MIT.EDU> you write:
>In article <7656@apple.Apple.Com> lsr@apple.UUCP (Larry Rosenstein) writes:
>>						       MultiFinder is not
>>large (50K) and it works very well.
>
>Well, size is all relative, you know.  50K is more RAM than my
>friend's Apple II ever had.  And almost as much as the roms in the
>orignal Macs.

50K includes both (C) code and data, as opposed to the ROM, which is code only.
This number would have been about 9-10K if it had been built in from the start.
However, as new ROMs get created the MF code can be put into ROM, just as
system patches are.

> But whatever your standards, you can't say 50K is
>insignificant. 
>

True, but hopefully what you get for 50K isn't either.  Besides, it definitely
pales in comparison to the size of kernels of other OS's, and only a fraction
of MultiFinder is its kernel.

>>>What's that?  Mac software?  Who cares about Mac software?
>>
>>Millions of people care.
>
>Sigh.  I forgot to put in the :-).  I know perfectly well that my
>argument is an attempt to compare multitasking and MultiFinder in a
>vacuum.  The logic is so clean when you ignore other variables.
>
>Reality is messy.  I know that millions of people own Mac software,
>and that this is MultiFinder's greatest (only?) advantage.  Of course,
>_I_ don't own any Mac software, so why should I care, right? :-) :-)
>
>If all you want to do is compare multitasking and Multifinder _as_
>_tools_for_switching_between_programs_, I stand by my argument.  If

I agree.  MultiFinder has made life more difficult (only a little, though)
for application developers in order to make life a great deal easier for
end-users.  I think much of the argument depends on what point of view
you take.  Adam argues from the programmer's point of view, Larry (and
other Apple people) from the user's.

>>If you simply judge systems on the basis of multitasking, then MultiFinder
>>would not offer any advantages.  Most users, however, don't buy computers
>>based on operating system features.
>
>Hear hear!  Couldn't have said it better myself!
>

Again, it depends on what point of view you take.  MF offers many advantages
for the user.  It offers most of the same advantages as preemptive multitasking
for well-behaved programs.

I think one point you are both missing out on is that preemptive multitasking
requires hardware support not available on a 68000.  It is not compatibility
that stops us from implementing preemption, although it does pose some hairy
problems, but the requirement that we can run on a MacPlus and SE.  It may be
that preemptive multitasking is preferable to cooperative multitasking -- this
is the real argument here, at least the really interesting one -- but it isn't
possible on a MacPlus.  We are investigating preemption on the MacII.

As far as which type of multitasking is preferable, I think the answer is, a
combination of both.  Even if preemption is available it is still *highly*
desirable to allow the application to notify the OS that it has nothing useful
to do for a given time period.

-Phil Goldman
Apple Computer