🚀 go-pugleaf

In article <danpop.824477998@rscernix>, danpop@mail.cern.ch (Dan Pop) wrote:

> In <4g00uj$gaq@nuke.csu.net> ghansen@isc.sjsu.edu (Gregory L Hansen) writes:
> 
> >In article <danpop.824304112@rscernix>, Dan Pop <danpop@mail.cern.ch> wrote:
> >
> >>My point was that even software written by Apple, even as part of MacOS,
> >>has problems with CMT, so there is little wonder that other applications
> >>have this kind of problems, as well.  It's easy to see why Apple is
> >>moving towards PMT, but it's impossible to understand why it took Apple
> >>so long. (Well, the reason is that the average Mac user thinks that
> >>this is how computers are supposed to work and accepts all the crap s/he
> >>gets.  The best example is offered by the Windows users, who accepted
> >>the frequent crashes for the same reason.)
> >
> >I agree it's easy to see why Apple is moving toward PMT, and I wish 
> >they'd started down that path years ago (although I didn't know the 
> >difference a year ago).  But I think it's also pretty easy to see why 
> >they didn't.  It involves rewriting the OS while maintaing backward 
> >compatibility.  You see how long it's taking to get Copland out.  If they 
> >had started this earlier, they probably wouldn't have finished before 
> >they switched to the PowerPC line and would have had to start over again 
> >anyway.  And if they had finished, what's the point of putting out new 
> >software technology for a line of hardware they're about to drop?  And if 
> >Copland was going to run on the PPCP, they sort of needed to finalize 
> >that standard before they could finish the new OS.
> 
> You mean that Apple programmers are so incompetent that they're not able
> to write code which can be easily migrated from one hardware platform
> to another?  Linux was ported to a bunch of platforms and Apple is porting
> it now to the PMac (they don't even seem to wait for the PPCP).  If this
> was possible (so easily) with Linux (or *BSD), why do you think it would
> be so difficult with software written by Apple's professional programmers.

Perhaps you've given us the answer right there. Linux is being ported for
3 models of Powermac. It won't work on the several dozen other models of
Powermac. So much for easy migration. Part of the problems in recoding
MacOS is that some parts are assembler written in 1983 by people that
aren't in the company any longer. That code works great and is rarely used
(like the floppy formatting code) so why fix it right away. It's not a
matter of recompiling or cranking out a new driver here or there, but
completely rewriting most of the OS. Ask the Win95 team how easy it is to
just push the button and have it work.

That early code was clearly not designed to be migrated. It was designed
to work quickly and take up little space on limited hardware. They
sacrificed portability for speed and size. What programmer hasn't done
this? Apple's programmers were so incompetant that they shipped a GUI
based OS on hardware so limiting that MS could only manage DOS. I wish I
could be so incompetant. Had Apple written nice portable code back in
'83-'87 then we wouldn't have a Mac to argue over. It's a Catch-22. Either
they do the wrong thing and have a product to fix, or they do the right
thing and go out of business. You pick.

> >I'm not saying that makes it all right, just that I'm not really 
> >surprised it took that long to get PMT on a Mac.  But I do wonder why 
> >Windows users haven't been able to enjoy PMT until the hybrid form of 
> >Win95 came out.  Microsoft doesn't exactly have to deal with radically 
> >changing hardware, you know.
> 
> Windows users had the choice between at least a dozen of OS's with PMT.
> They could even run their Windows apps in a PMT environment, at least on
> OS/2 and SCO.  The PMac users had MacOS as their only option.  Can you see
> the difference?

Sure. But since the *vast* majority of Intel users run Windows and not all
of those other things I wonder what the value of that statement is. Either


a) Wintel people, by and large, have been mistakenly pouring resources
into Windows when some PMT OS should have been selected instead.

b) Most people really don't care about PMT. It doesn't affect their life
that dramatically.

Personally, I vote for b. I don't think that 100 million Windows users are
that stupid. Sure there are people on the Mac that would like PMT and have
been denied. Nobody is arguing that. But have *most* people on the Mac
been disappointed by Apple by not offering a PMT environment? No more than
*most* people using computers today have been disappointed. Most people
simply don't care. You and I do, sure. I'm willing to let PMT go for the
other features that the MacOS has. PMT matters to me, but not as much as
other things. It's a trade-off.

I'd rather have Applescript than PMT. I'd rather have Appleguide than PMT.
I'd rather have OpenTransport than PMT. You see 2 reasons why I use a Mac
and one reason why I'm not disappointed by Apple not shipping a PMT OS by
now. I could go on and on but perhaps you see my point.

> Moreover, with SCO I was even able to run Windows 3.1 apps from my
> non-intel workstation by telneting to the SCO box and starting the
> Windows emulator there.  The SCO emulator is X-based, so I could get the
> emulator window on my remote screen without any problems.

Great. If I or anybody else I know ever want to do that on a Mac I'll be
sure to send the envy vibes your way. Sure it would be nice to have a tool
like that but I don't know many people that really need it. We have
Timbuktu that runs on Mac and PC and does a similar thing (though only
from a Mac or PC with Timbuktu) and that seems to satisfy most of that
group. How many hundreds of Mac users would want that feature of SCO? How
much would that cost again?

> >One thing I don't know the answer to, though, is why the CMT System 7 
> >makes users wait for printing and disk formatting.  I'm sure it didn't 
> >have to.  Luckily I can print in the background and I rarely need to 
> >format floppies.  But it's still kind of annoying when everything stops 
> >if I click on a menu or on a window title bar or something.  
>
> Apple programmers didn't try very hard to get their software right.
> If it was good enough for the average Mac user, it was good enough for
> them, too.  The current state of MacOS (especially the idiocies and
> machine code inherited from the 68K version) shows this very well.

Understood. There is a bit of a problem with Apple's mentality, I'll
admit. But it is no different than the mentality shared by every other
company out there. Does MS not ship when NT is good enough for the average
NT user or SunOS is good enough for the average SunOS user? That's
economics, bud. That's how it works. You don't see Mac people complaining
*that much* about the MacOS, only NT people or Linux people. You're
arguing that Apple should make the MacOS good enough for the average NT
user or the average OS/2 user. 

But there is merit in that argument. *You* will never be a strong Mac user
unless they take that approach. Of course if that means abandoning many of
the things that Mac people like, that's not good either. Apple, like
Microsoft with Windows, needs to strike a balance between advancing the OS
and strengthening the OS. Some OS's primarily strengthen, the UI issues
are secondary (UNIX, for example).  That's fine for many people, but not
for most. Some OS's primarily address UI issues, and strenthen secondary
(Win and MacOS). That's what most people seem to want (whether or not we
agree with them) and that's what the companies need to deliver. Do I want
PMT? Sure. Do I want the best UI more? Yep. Am I willing to trade PMT for
a good UI? You bet. I have 6 people here. I have crashes about 1x/week
through the whole office. I can handle that easily. If I gave these people
OS/2 or NT, I'd be hand-holding most of my users through the simplest of
tasks - something I don't have time to do and something that isn't a
problem on the Mac.

Does the criticism of the MacOS apply to NT and OS/2 as well? After all,
they may now have fully native support for the PPC but they sure as hell
didn't have a product 2 years ago. They struck a balance as well. A
different balance than Apple did, sure. They decided that nothing for a
year or two was better than something for 3 years. Oh well, I guess when
your potential installed base is as low as those two OS's on the PPC you
can make those decisions. When it is as high as the MacOS (what, 5 million
PowerMac users now) then you have to make a different decision. Oh, and
how many OS's don't have idiocies? 

I have to agree in part with your original statement as to why it took so
long to get PMT. I think it should have been implemented around the time
of the MacII or so, when the PMMU was first introduced to the Mac line.
There wasn't too much backward support needed and compared to what needs
to be done today it would have been a *much* easier task. 

I can also understand why they didn't do it. At the time they were
introducing at most two machines per year. At that rate they would have
introduced 14 machines, not the 120 that actually occurred. The system
software changes were pretty mild until 6.x and 7.x. Apple must have
thought they had plenty of time to make the change. They didn't forsee
what would really happen. Whether Apple blundered or simply got caught in
an unforseeable situation can never be resolved but it happened. It
happened to MS as well. However, through good acquisitions MS was able to
offer a more robust product (NT). Apple has not decided to make the same
move, perhaps foolishly. 

Keep that in mind, though. The top *three* OS's (DOS, Win, and MacOS) all
suffered from all of the things you mention for years and years. Perhaps
Apple and MS know a thing or two that you don't...

-- 
Bob Cassidy
UC Irvine