26 Jan
2002
26 Jan
'02
3:25 p.m.
I don't know where you were buying DRAMs for that kind of money. I was
working a baseband processor design in '87 (VERY cost-sensitif due to volume)
and was constantly seeing prices of about $1 for 41256's at the time. I heard
rumors of price surges, but they didn't show up in distributor prices, and
certainly not in direct pricing from the vendors.
If you didn't mind the teensy hard disk, you might get by at $30K, but if you
wanted it equipped with a video display capable of 1024x768, 24 MB of RAM, a
pair of 750 MB hard disks, you were looking at a MUCH bigger price tag, and
particularly if you wanted to use a DEC monitor.
Then there was the networking hardware. If you wanted to be compatible with
non-DEC stuff, you had to "dig deeeep" as the preacher used to say ...
more below ...
Dick
----- Original Message -----
From: "Ethan Dicks" <erd_6502(a)yahoo.com
To: <classiccmp(a)classiccmp.org
Sent: Saturday, January 26, 2002 7:04 PM
Subject: Re: Bell & Howell Apple II update
--- John Allain <allain(a)panix.com> wrote:
Terminals? Who wanted to use terminals? Nevertheless, with an add-on board
costing a fraction of what a DEC terminal cost, you could add a 16-port board
to a PC/AT along with the software. Now, I don't know what you'd do with such
a thing.
- - - -point
Just compare the cost and features of the
PC/AT clones sold in, say, '87-88 with a similarly equipped microVAX-II.
The PC/AT would typically cost about $800 bucks, while a similarly
equipped uVaxII cost nearly $100K...
$100K? An 11/750 in 1982 cost that. In 1986, when they were first
out, a uVAX-II was about $30K for a standard system - 9MB of RAM and
a 73MB MFM drive. Yes the drives from DEC were expensive. They always
were poor at a price/capacity ratio, especially compared to third-parties.
That $800 AT may have had a larger disk, but it sure didn't have much
RAM. 41256 RAM chips in the tube (not assembled into boards) shot from
$3.50 each to $17.50 *each* in 1987 due to the U.S. Dept of Commerce
imposing dumping fees on Korean DRAM vendors. I was trying to populate
a Spirit Inboard (1.5MB) on my Amiga - 1MB x 9bits (parity) was going
for literally $630 per megabyte.
Also, the AT was old tech in 1987 and the uVAX-II was a year old. That
AT was $5K a couple years earlier when _it_ first came out - with 1Mb of
RAM and a 40Mb ST-251 drive.
Inside a year, the power cost alone exceeded the
PC/AT...
8A max draw for a BA-23 is 880 watts. At $0.08 / KWH (our rates in
Columbus), that's $0.07/hour in electricity, or $1.68/day or $613/year...
less than the cost of 1MB of RAM chips.
yet folks LOVED
the microVax and hated the PC/AT clone, that ran half-again as fast.
How many terminals could you put on that stock PC/AT clone? The regular
complement on a uVAX-II was 5 - console plus 4 DZV11 ports. How many
people could use the PC/AT clone at the same time? 1? Don't forget
that a *big* chunk of the price from DEC was the software license. If
you needed VMS, you *needed* VMS. The uVAX was near the *bottom* end of a
line of expensive machines and could nearly match integer (not I/O)
performance of an 11/780 that cost 10 times as much ($300K new in 1978 vs
$30K new in 1986).
That's another point - how fast could you pump bytes through a PC? The
ISA bus was 8MHz - and that doesn't mean 8 million 16-bit words per second.
The Qbus in a uVAX can pump over 1MB/sec, but I don't have the numbers
here in front of me.
It can, in theory, but, having had to do performance testing, I'd say that on
a good day, with a tailwind, it will do 200 KBps. The ISA bus was operated
from an 8MHz bus clock, but it used two clock ticks per
bus cycle, and
typically one for an idle cycle if it had to go back to the bus for
anyting.
It didn't typically use the bus for memory, however, unless you had a big RAM
board plugged in, though with linear addressing that was limited to 16 MB and
couldn't practically use all that either. Normally one split a big RAM board,
like my 24MB board between "extended" and "expanded" (banked) memory.
In any
case, the ISA couldn't really transfer much more than about 200 KBps, either,
when doing useful work.
Apples and oranges. Comparing a PC-AT and a uVAX-II - Yes... the PC clone
can add numbers faster, but there's lots of things it couldn't do, things
people with MicroVAXen *wanted* to do (1 machine for e-mail for the entire
department, as a personal example).
Fact is, the PC/AT can compute faster. The uVax-II had a more flexible and
more secure OS, though. However, it's silly to compare a computer costing
many tens of thousands of dollars with one costing a few hundred when the
cheaper one has more MIPS, more STORAGE, and, in fact, costs less per user
just on the basis of the cost of a terminal, which nobody wanted anymore by
then. Fact is that DEC's gone now because they took the wrong approach to
marketing their hardware. They made the stuff slower, larger, and VASTLY more
costly because it gave them some quick profit. That didn't provide them a
future, though. Others have yet to learn that same lesson.
It's that notion of using a single machine for email for the entire department
that killed DEC, and other dinosaurs like it.
I'm glad they're gone because they made their money by making computing
expensive and the domain of the very rich, while they could have made a lot
more money by making their computers cheaper, more avaialble, and more
compatible with the rest of the world. Hindsight's always 20/20, though.