🚀 go-pugleaf

In article <40nrjt$q9g@eken.hv.se>, henke@hv.se wrote:

> In article <rmcassid-0808951720080001@dante.eng.uci.edu>,
rmcassid@uci.edu (Robert Cassidy) writes:

> :But how much more flexible is UNIX over a Mac? A list, perhaps?
> 
> Memoryprotection and premptive multitasking gives each process an illusion of
> running indepent of all the other. A processes never have to care about the
> other. Therefor is the knowledge limited for each process. A system
where you limit
> the knowledge to only those who need it, it wellknown to be much more flexible
> and expandible.
> With cooperative multitasking do every process have to think of all the
> others and leave the CPU by themselves.
> Now, which one do you think is most flexible by definition, right from start?
> To start with a list with things a flexible mac should handle, and unix
already does:
> * implement the OS on, let's say, 128 * PowerPC 601.
> * change window manager.
> * caching ethernet network data, to decrease network load.
> * share one machine between 100 remote users at the same time.

> Come on, you can't be serious asking this question? It isn't fair to compare
> toys with professional tools...

But the things that you have mentioned are used by a relatively small
group of people. I'll agree (except changing the window mgr. and caching
ethernet network data with OpenTransport upon us) that the MacOS is
limited in the ways you have described. How many people have 128 CPU's to
play with? How many people need 100 people leeching off of their machine?
Those people that do use UNIX but I think that those numbers are below the
level of millions and into the 10K or 100K range. MacOS is in the 20
Million range right now. For the majority of people, UNIX's only benefits
are PM's and those aren't very strong benefits to most people - they are
only improvements over what the Mac already has.

> :Then it was not configured properly - same argument as a UNIX system. I've
> :never had MacTCP crash. It's all configuration.
> 
> Please understand... MacOS have no memoryprotection, therefor can a program
> go nuts and destroy whatevver it feels like with out any limitations.
> Sure, nobody wants this and does everything to avoid it, but you can never
> prove that any software is bugfree (if you can, would for sure thousands
> of software producing companies like to hear how you did it).
> Unix on the other hand, accepted that software might contain bugs, but
> concentrated and limiting the effect of them by using memory protection.
> Just like information hiding is one of the strongest features of
> object oriented programming hides data from each other, does MP hide
> processes from each other.
> You only have took at the basic architecture of MacOS and Unix to understand
> which one is the better.

I'm not arguing that UNIX is not better designed from the standpoint of
memory or multitasking. But a UNIX box with improperly configured TCP may
not crash, it will simply be unusable as 2 of 5 points listed above depend
on TCP. What difference if it crashes or is unusable? The point is that
both systems need to be configured properly.

As for the 'go nuts and destroy whatever it feels like without
limitations' - would that be the same as saying that without marshal law a
society could 'go nuts and destroy whatever it feels like without
limitations'? This society operates cooperatively and with minimal
protection - not that the analogy holds completely - but reliability is
relative to conditions. The conditions under which the Mac tends to
operate can tolerate a lower level of reliability. All this will change
soon, though.

But keep in mind, there are limitations on the Mac and that the argument
only holds if it's routine that software 'goes nuts and destroys whatever
it feels like'. That really doesn't seem to stand up in practice. My Mac
crashes about as often as the campus' UNIX mail gateway goes to hell - I
don't see the argument.

> :And most Mac users know what system does what they want to do with the
> :least pain. I also know what system has the most potential to support
> :future developments! (Primarily closed systems and systemwide support of
> :technologies reduces redundent development and increases
> :interoperability).
> 
> That's what Mac people often believe. It's ok for them to think so, it is just
> that it's annoying when they try to convince people that a Mac Quadra
should be
> better than any Unix workstation, no matter the task is... It's like ignoring
> the last decades discoveries in computer science.
> Most mac people still think Unix is the same old system as it was in the 70's,
> but things have changed. Unix is becomming a lot better each day. The
complexity
> of Unix is decreasing. It is no longer so difficult to learn is many people
> think it is. With today's powerful computers is it avaible for most people,
> so it DOES compete with MacOS.

I agree that that is annoying as well. I'll be the first to admit that
UNIX is the right tool for a lot of tasks. I'll admit that DOS is the
right tool for a lot of tasks. But for most tasks that most people do,
I'll stand by my claim that the MacOS is the right tool. 

I'll also continue to argue that UNIX does not compete with MacOS. As it
seems pretty clear that even MacOS is too complex for most people, I'd
have to say that in spite of UNIX's advances in ease-of-use, it is still
way out of most peoples reach.

With the tasks that have been brought to the average user because of the
Mac - desktop publishing, desktop video, etc. etc. I'd have to say that
the UNIX is still missing the majority of the market that Mac and Windows
is reaching. In spite of being free and open and protected and preemptive
and all of those things, UNIX is not being chosen by people in any
significant sense.

-- 
Bob Cassidy
UC Irvine