🚀 go-pugleaf

In article <dke-2203960645550001@adnline48.adnc.com>,
David K. Every <dke@adnc.com> wrote:
>In article <root.827273089@rscernix>, danpop@mail.cern.ch (Dan Pop) wrote:
>
>> >That's your problem. If the memory model was the pinnacle of the 
>> >usefulness of an OS we'd all be using Unix. I wonder why we're not?
>> 
>> Mostly because you don't have the necessary skills :-)  And because
>> Microsoft Office doesn't run (natively) under Unix :-)
>
>That is sorta the point. 
>
>You are wrong (specifically), I have all the skills to be able to use
>Unix, and have used it - VMS, RDos, Arts, Pick, Dos, etc. But most users
>do not - nor should they waste their time learning to interact with a
>computer in a less efficient way (for all but the top 1%). The computer
>should use ITS resources to make it easier for people to do THEIR work!
>That is the whole point of HAVING a computer!

Certainly the point of having a computer is to make your work more
efficient than it would be without the computer (and, indeed, to
maximize that efficiency).  However, you make a common mistake with
your statement above: you equate ease of use with efficiency of use.
These are demonstrably not the same thing.

>That is WHY CLI's have been failing. Most users recognize that altruism in
>that statement. Many CLI users do not!

CLIs have been failing because they are harder to *learn*, not
necessarily because they are harder to *use*.  However, since the CLI
in question is almost certainly the one supplied by DOS, this amounts
to basically the same thing.  The DOS CLI isn't even in the same
league as the ones supplied with Unix, but since the CLI in DOS is the
most common one (thanks again to Microsoft, the company that has made
an art form of making mediocrity successful).

Either way, a CLI is merely one way of using a computer, and while you
and many others may argue that it is an obsolete interface, it should
be clear to anyone who has thought this through that a CLI and a GUI
are complementary.  For simply launching programs, I find the GUI to
be a better tool than a CLI...as long as the variety of programs I
want to launch doesn't grow too large (otherwise I have to look for
them or otherwise select them, which takes about as much time as does
typing in the name).

>> Yes, but it's memory protection and multitasking which are currently the
>> limiting factors in MacOS and Windows.  This is why it makes a lot of
>> sense to talk about how they're handled by various versions of these 
>> popular OS's.
>
>Except that users do not care about specifics again - nor should
>they.

They shouldn't care about the details to the point that the details
don't make any difference.  But as with many things, the details make
quite a bit of difference unless what you're trying to do is the same
as what everyone else wants to do.  Since a computer is essentially a
general-purpose tool (because it's programmable), it's rare that you
find people who want to do the *same thing* as everyone else is doing.
Instead, each person has his own needs, many of which will overlap
with the needs of others, many of which are (relatively) unique.  A
single computer can meet most, if not all, of those needs, but the OS
has to be up to the task, and many aren't.

>If your car has double wishbone suspension or fully independant is
>irrelevant in 99% of the cases - what you care about is HANDLING! 

The problem is that 'handling' is a very broad brush.  *How* do you
want your car to handle?  Do you prefer to be able to steer it with
the throttle in the corners?  Do you prefer it to lose traction in the
front before the rear (i.e., to understeer)?  Do you prefer it to
oversteer?  Do you prefer a neutral car that you can simply drift
through the twists in the road?  Do you want to be able to tune the
car's suspension for different situations or do you prefer it to be
fixed (i.e., an overall compromise)?

>Wether
>it is a diesel, wankel or traditional is also irrelevant - you care more
>if you can get the fuel you need cheaply and readily, what its milage is,
>and its maintainability/performance.

But the problem is that each type of engine provides a different *set*
of those characteristics.  None is better than the others in all
categories.

>Now apply that to computers. Full memory protection and PMT are good
>things - but they are not the total of the car. The fact that Unix is used
>so little (despite having those things) and DOS sold so well (despite NOT
>having those things) is proof that most users DO NOT CARE, or are not
>willing to pay the price (at that time).

You forget that there are two sides of a market, the supplier and the
consumer.  The consumer is going to be limited to what is available,
and it is up to the suppliers to supply something better.  I think it
is more the fault of the suppliers that memory protection and PMT were
not made available to the consumers.  When done right, they provide a
very noticeable benefit, and as I've argued many times in the past,
they do not require significant computing resources.  The computing
power at the time Microsoft released Windows 3.1 was easily sufficient
to support PMT and protection, and Microsoft could easily have
implemented both at the time had Windows itself not been such a badly
designed piece of junk to begin with.

>People want stability, but they are not always willing to sacrafice the
>other costs. (Performance drops, memory penalties, etc.). 

The problem is that you incorrectly believe that performance and
stability are significantly sacrificed to get memory protection and
PMT.  This is not so, and demonstrably so.

Consider the Amiga.  It provides PMT *and* a GUI on a 256k system.  It
also happens to provide shared libraries, loadable filesystems,
loadable device drivers, and a CLI (and can do all that in 256k).

Consider Linux.  In its early incarnations, it would run on a 1M 386
system.  It would have required less had it not been implementing
networking, multiple filesystem support, terminal handling, etc., all
in RAM (the Amiga has the advantage that the code can be in ROM.
Linux doesn't have this advantage).

>Now that resource costs are coming down. And we can pay for the lost MIPS,
>and add the extra memory, they are becoming viable for home/buisiness
>computers. But they were NOT viable (or worth their costs) for MOST users
>for the past 10 years! The reason why Unix and NT (combined) have been
>relegated to specialty tasks, and less than a small percentage of apples
>marketshare .  NOW it is becoming viable - and guess what, the MacOS is
>getting those features. Big suprise. 

The reason MacOS is now getting those features is that the consumers
*finally* have a different choice that provides those features, and
those who have experienced the stability provided by memory protection
and the performance of a decent PMT system are now demanding these
features in what they buy.

Again, the *suppliers* must advance the state of the art.  The
consumers can't because they can only compare what's available.

>In other words MacOS is being UserDriven... Unix and NT where technology
>(engineer) driven. Which is more in the users best interests?!?

The answer to this one should be obvious.  A system which is
technology driven (as long as it doesn't sacrifice user-friendliness
in the process) will be in the users best interests because it is
*more* capable than what the user needs at the time he buys it...and
this means that the system will be able to deal with the *future*
needs of the user as well as the current needs.

If you, like Microsoft, believe that a user should replace his
computer, his OS, and maybe even his software once every couple of
years, then you may have a point.  But even that isn't in the best
interests of the user because it's more expensive for him to do that
than it is for him to get the most capable system he can.

And for the record, most user interfaces available for Unix don't cut
it as far as I'm concerned.  Even CDE isn't very well engineered (what
I've seen of it, anyway)...just pretty.  However, everything I've
heard and seen says that NeXTStep is very, very good...better than the
Mac GUI, even.

I'm really looking forward to the release of OpenStep and, finally,
GNUStep.


-- 
Kevin Brown                                     kevin@frobozz.sccsi.com
This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end >
            This is your .signature virus on drugs: <>
                        Any questions?