Path: utzoo!attcan!uunet!unsvax!jimi!tahoe!apple!das
From: das@Apple.COM (David Shayer)
Newsgroups: comp.sys.mac.misc
Subject: Re: The New Macs: Greedy Compromises?
Message-ID: <47012@apple.Apple.COM>
Date: 2 Dec 90 06:43:58 GMT
References: <1990Nov30.205419.14100@grape.ecs.clarkson.edu>
Organization: The Troll Den
Lines: 98

In article <1990Nov30.205419.14100@grape.ecs.clarkson.edu> anthonjw@clutx.clarkson.edu (Jason W. Anthony) writes:
>>>"OK the LC has at 68020 so that puts it in line with Mac II power, but
>>>oh, is has 16-bit pathway instead of 32-but, but it does has 32-bit to
>>>the ROM."
>>Anyway, why do you care?  The user has no clue how wide the data path is,
>Because now when I try to compare the speed between two machines, it becomes
>much more difficult.  Which is faster, the Mac II (16 MHz, 68020) or the
>LC (16 MHz, 68020)?  I can look at benchmarks and see the Mac II generally
>is, but how will it affect the specific things I may want to do.  Will the
>smaller path not be much influence, or will the programs I want to use
>happen to make it a significant bottleneck.  For example, take the article
>in MacWorld on the LC (December 1990).  On page 187 it shows two benchmarks,
>one for "Fit Image to Window" and one for "Preview Image".  According to
>MacWorld, they both excersize the processor similarily, but the first
>test was faster on the Mac II, and the second faster on the LC.

If one test is faster on one machine, and the other test is faster on
the other machine, my conclusion would be that MacWorld is wrong when
they say that both tests exercise the processor similarly.  

But the real test is, is the machine fast enough for you, when you do
your work on it?

>>Graphics go faster with on board video, not Nubus video, because the
>>processor can access the video RAM directly, instead of going through
>>the 10MHz Nubus controller.
>Yes, but the keyword is video RAM.  The 68030 is slowed down because it often
>has to wait to use main memory because the Memory Decoder Unit (MDU) is
>accessing the same memory while creating images on the screen.  There is
>no seperate "video RAM".  (This is on page 193 of the MacWorld article.)

Yes, video access is faster with on board video, while non-video
computing is slower.  You have to decide if you want faster number
crunching or faster redraw.  I mainly compile, so I want faster
processing, thus I use a nubus video card on my Mac IIci.  I know people
who want faster redraw time, so they use the on-board video.

But I think this type of decision is beyond what most customers care
to learn about the computer.

>>Most users don't do anything that requires an FPU, so its not
>>included standard, to save the average user money.  You can add one
>>if you run something like Mathematica (on a IIsi, God help you).
>>Third parties are going to sell FPU boards for the LC.
>Yes, but it used to be that if you had a 68030 you knew you also had
>a FPU.  Now this is another one of those "little things" you have to
>remind yourself of.

Again, I disagree completely.  Apple has been putting out tech notes for
a long time saying don't assume that if the machine has one feature
(a 68030) it will have another feature (a 68882).  Instead, use the
system calls to check for the exact feature you want.

>I'm sure too.  But now new users buying an LC have to deal with different
>resolutions.  Why did Apple make the lower-resolution 12" monitor?  (I know,
>money.)  Try helping a computer novice pick between 512x384 pixels and
>640x480 pixels.  Their first question: "What's a pixel?"  Why didn't Apple
>lower the price of the 640x480 a little more, instead of putting in time
>and resources (== $$) into making a new one?  

Helping someone pick between the two monitors is easy.  You show them
how much text shows up in MacWrite on one screen, and show them how
much more shows up on the other screen.  Then you tell them the price
difference and let them decide if the extra screen size is worth the
money.

>Nope.  But why 1 Meg in the Classic, and 2 Meg in the LC, and none in
>the IIsi?  Again, it makes the buyer's life a little more complicated.
How does it make the buyer's life more complicated?  They tell the
dealer how much memory they would like in their new Mac, and the
dealer puts it in before they ever see the machine.  That's what we
did in the dealership I worked in.  We didn't worry the customer with
what Apple's standard confidurations were, we asked them what they
wanted, and put that system together for them.

>Exactly!  Yes!  Don't you see this is beginning to happen to you!  These
>new machines are creeping in all these little exceptions and things
>to be careful of.

I think we have a couple of basic disagreements here.   First, I don't
think an average buyer should buy a computer based on technical specs
like MHz, MIPS, processor type, data path width, on-board vs nubus video,
or any of that crap.  They should buy a machine that has things they
care about.  They should decide if they need a color screen, and how
large a screen.  They should try the software they plan on using, and
buy a machine that's fast enough for them when it runs their software.
Finally, they need to balance what they want with what they can pay for.

Second, I think its inevitable that the Mac line become, in your words,
more complicated.  Apple has to bring out new machines, and the new
machine have to have more features, or be different in some way, or else
why bother to bring them out at all.  Thus you inevitably get a bunch
of machine with slightly different features.  But I think this is good,
it gives the user a large range of choices.  I think you are advocating
stagnation.

David
No disclaimer- screw the lawyers.