November 1997 feature by Steve Hawley |
||||||||||||||||||||||||||||||||||
|
Occupied?
I have a theory. Actually, I have quite a few theories, like the
one about how the tape player in my friend David's Volvo causes it
to rain in Ohio, but those are not salient. No, the theory I
will expound upon is the cause of Apple's tortuous demise.
The cause are the numbers 19136 through 19199, inclusive. These
numbers represent the opcodes for the instruction Test And Set (TAS)
on the Motorola 68000 family of microprocessors. TAS does two
things: it reads a byte in memory and checks to see if it is
non-zero or negative, then it immediately sets the byte to -127.
The important thing is that these two operations happen in one
instruction.
"So what?" I hear you cry. Well, think of it this way: Test
And Set is the occupied light and the latch of an airplane john.
Normally, you check to see that the light is off before you
enter the john, and if it's off, you step in, flip the latch
and go about your business. If these actions were separate,
the following scenario would be possible: you check the
light, step in and before you reach the latch, another
passenger checks the light, steps in and flips the latch.
Disaster has struck: both of you are in the john, voiding yourselves
all over each other.
The hardware designers of the Mac, the same ones who put their
autographs on the insides of the case, chose not to put in correct
hardware support for TAS.
With TAS it is possible to write a multitasking operating system.
Without TAS you end up with programs with the ability to urinate
all over each other.
What was worse was that even when Apple had corrected this in
hardware, they couldn't correct it in software because of all
the legacy machines that wouldn't be able to work with it.
Had Apple chosen to risk alienating customers early on by
orphaning early machines, they could've had machines that
were full multitasking, more responsive, and more robust.
Now, it is too little and far too late.
I sit and type on a PowerBook 520c. It's a nice little machine,
but Apple in its infinite wisdom, committed a much more complicated
hardware error so that this machine, like those without TAS, is
also dead-ended.
While I can get a hunk of silicon, epoxy and metal which will
transform this laptop into a PowerPC based machine, I have to
ask myself the question "Is it really worth the money?"
The answer is a resounding "no". This machine, although it
is a scant few years old, is on its second back hatch to cover
the ports (and this one doesn't close too well); is on its second
bezel that covers the microphone and cables to the LCD screen
(and this one is breaking in the same way as the last);
is missing the hook that latches the case shut (sheared off);
is on its third complete formatting of the internal hard drive;
has serious stress cracks on the cover near the hinges; and has
serious stress cracks at the dividing line between the keyboard
and the battery compartment.
I was not a harsh user of this machine. Really.
So a third party, for $500 will sell me a little interface board
that will cost me $75 to have a dealer install into a disintegrating
case so that ultimately I still can't run the current revision
of the operating system from Apple, all because of hardware mistakes.
When people ask me "what kind of computer should I buy?" the
answer I usually give is, "the one that runs the software you want."
I usually add the side comment, "and if you can arrange it, get
your company to buy it for you".
I will certainly follow my own advice if I choose to replace
my PowerBook, and because the software I want to run is less
frequently available on Macintosh machines, I seriously doubt
that I would ever buy a Macintosh again.
in the junk drawer:
October 1997
|
|
