Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!ucla-cs!lanai.cs.ucla.edu!lange
From: lange@lanai.cs.ucla.edu (Trent Lange)
Newsgroups: comp.sys.mac.misc
Subject: Re: The New Macs: Greedy Compromises?
Message-ID: <1990Dec5.113315.9159@cs.ucla.edu>
Date: 5 Dec 90 11:33:15 GMT
References: <109755@convex.convex.com> <1990Dec4.092217.26859@cs.ucla.edu> <109822@convex.convex.com>
Sender: news@cs.ucla.edu (Mr. News)
Organization: UCLA Artificial Intelligence Laboratory
Lines: 163
Nntp-Posting-Host: lanai.cs.ucla.edu

I certainly don't want to get into an ongoing thread about this, but...

In article <109822@convex.convex.com> ewright@convex.com (Edward V. Wright) writes:
>In article <1990Dec4.092217.26859@cs.ucla.edu> lange@lanai.cs.ucla.edu (Trent Lange) writes:
>
>>This is the second time in as many days that this little bit of misinformation
>>has been spread, so it's time to correct it.
>>
>>- *All* NeXT machines come with floppies.
>
>Funny, I have a brochure for the original NeXT cube -- still being sold
>at discount prices -- right in front of me and the only removeable media
>it mentions is the optical disk.

Funny, I have a brochure on the original 128K Mac -- still being sold
used at discount prices -- and it only has 128K.

NeXT is not making or selling the original NeXT cube anymore.  They
were selling a few leftovers from their warehouses at good prices for
a while, but they only sold them in systems that included 68040 upgrades
and an external floppy.

How Businessland sells their few remaining old cubes is another story.
However, I'd hardly expect somebody to fault Apple for selling Macs with
non-standard 800K drives now simply because some stores are still getting
rid of their last Mac Pluses.

>>As for all those other "nasty choices" (color and whether to get an
>>optical), the average user doesn't have to worry about them.  The
>>NeXTDimension color is a production-quality 32-bit board, coprocessor,
>>and monitor system that is priced accordingly.  
>
>So NeXT eliminates those nasty choices by making them so expensive no
>one can afford them, huh?  Well, at $20 million a pop, the average user
>doesn't have to worry about buying a Cray, either.  I'm sorry you feel
>Apple is offering the users too many choices.  Some people *like* having
>this flexibility.  Yesterday, I had to make the choice of whether to
>continue using the 8-bit internal video on my Mac IIsi or buy a 24-bit
>RasterOps graphics board for $399.

I'm sorry that I didn't explain this in detail originally, but I was
trying to avoid being tedious.  Since you took that as an opportunity
to attack, I'll explain it here as quickly as possible:  The NeXTDimension
Color is a production-quality 32-bit board with a 80 MFLOP i860 coprocessor,
a JPEG video compression chip (allowing playing of full-motion video
from a hard disk), and a 1120x832 pixel 16" color monitor.  The average
user doesn't need this.  However, if he *does* need it, he can get
it for his NeXTCube for $5100 (at educational prices).  The total price
for the NeXTCube and the NeXTDimension is less than the total price
for the relatively anemic IIfx, 8-24gc graphics coprocessor, and 16+"
color monitor, which anybody that needs the performance of the
NeXTDimension board will surely have to buy.

*However*, if the user is a mere mortal for which 16-bit, 1120x832 16"
color monitor will do, then he only has to choose a NeXTStation Color
machine for about $5200, and have a far more powerful machine and graphics
than, for instance, a Mac IIci for nearly the same price with a hard disk and
a 13" color monitor.


>>The fact of the matter is that NeXT has stuck pretty well to its
>>guns on the issue of keeping the lowest common denominator at
>>a high level.  All pieces of software written for the NeXT can
>>count on having a floating-point processor, at least 8 megs of
>>RAM, a DSP chip, unix, protected virtual memory, display postscript,
>>ethernet, and a screen with at least 1120x832 resolution.
>>So can NeXT users.  Software that can count on using more advanced
>>features can take more liberties and therefore do more things.
>
>There is nothing stopping you from writing a program that will only
>run on a Mac IIfx.  But doing that will limit your market to the small
>number of users who can afford a IIfx.  Similiarly, writing for the
>NeXT limits you to writing for users who can afford a NeXT.  The NeXT
>station defines the *highest* possible denominator in computing; there
>is (unfortunately) no such thing as a low-end NeXT. 

I assume you've seen the prices lately;  the lowest cost NeXTStation
costs about $3200.  You're right, the Classic costs less than that.
However, anybody that can afford a IIsi can afford a NeXTStation.
And, like you said, with that he gets the highest possible common
denominator in computing.

Notice the trend here:  two years ago, the original NeXT cube cost
over $7000.  Now it is less than half that.  And yet the NeXT's common
denominator has stayed the same (and risen in the current machines
with the 68040).

Five years from now, when the newest NeXT machine costs $1500, and is
within the reach of most users, their common denominator will *still*
have the same high level that there is today, because NeXT never
went and decided to sell 030 machines without a FPU or 68020 and
68000 machines that can't handle virtual memory but which the software
houses must cater to by neutering their programs.

NeXT's high common denominator is why Lotus is able to come out with a
highly-innovative program like Improv that requires high processor
performance:  they know that all the NeXT machines can handle it.

But pity the poor sap who goes out and spends $3000 on a Mac LC system,
and later decides that he wants to run Unix or Mathematica (decently)
or Improv-for-the-fx.  Yes, he *might* be able to get an upgrade
from Apple at their normal turnip-squeezing prices, but he better not
get a third-party upgrade, because Apple won't support it.


>>a few dollars more here and there could have made a big difference in
>>consistency and lessening consumer confusion (e.g. raising the price of
>>the LC by $25 to pay for the added VRAM needed to support the 13" color
>>monitor fully).
>
>First of all, I don't think VRAM chips are quite that cheap, but I could
>be wrong on that.  While we're at it, though, why not add another $25 worth
>of hardware to support sound input in stereo?  And since we've done that,
>another $50 for 16-bit CD-quality sound.  Changing the data bus to 32 bits
>surely wouldn't cost more than $100.  Neither would adding an FPU.  Of course,
>we want to use the 68030 to support virtual memory.  Adding a second slot
>probably wouldn't be that expensive.  And since we've already raised the
>price by several hundred dollars, a few hundred more would buy us 24-bit
>video.... You can always add features to any product, but you've got to
>stop somewhere.

You're right, but there are some things that are so commonly used
that it *is* worth it to include.  For instance, people tend to use a
SCSI interface, so even Apple makes one standard.  People tend to need
the equivalent of an 4-8 display card, so Apple builds one in to the IIsi
and the IIci.  They do that because simplicity of building it in to the
machine and the cost benefits of mass production make it much less costly
than they would be as an upgrade or a card.

This would seem to be the case with the added VRAM needed to support
8-bit color or grayscale on the 12" b/w or 13" color monitors.  *Lots*
of people use those monitors (even more so if the 12" color monitor is
as bad as it sounds).  Apple has *always* provided 8-bit color for their
color monitors.  And yet everybody buying an LC is going to go home, plug
in their 13" monitors, and wonder why they "can't get very many colors".
(Oh sorry, they can all depend on their competent salesmen to tell them
everything.  Right.).  This one is almost petty on Apple's part.

I won't argue the merits of all of the other little corners and feature
cuts that Apple has made.  Other somewhat disillusioned people here have
done so.  All I know is that NeXT manages to sell a machine for the price
of a IIsi that doesn't include any of its design compromises and has even
more hardware and software features than the twice-as-expensive IIfx.

So basically, yes, I think that Apple should spend the $100 or so extra
dollars necessary to have enough RAM to be useable, to have as many
machines as possible be able to run virtual memory, and in general
have a consistent and solid common denominator that both software
developers and customers can count on.  Or at least limit it to two tiers
(68000 and II/SE 30/LC class machines) rather than the four+ there are
now.  Even if it means having to have only a 45% profit margin rather than
a 50% profit margin.  The name of the game is long-term market share, and
Apple is barely starting to realize this.  The hodge-podge of corner-cutting
found on the new machines is what you get when they try to do it without
getting their feet wet, rather than trying to give more value for the
money.

- Trent Lange

-- 
************************************************************************
*      UCLA:  1990 NCAA Football Champions (yes, the other kind).      *
************************************************************************