 The magazine of the Melbourne PC User Group Age Before Beauty Tom Coleman |
 |
I was gazing gloomily into my poor old XT just the other day. Poor old thing! I have had good service out of it. I still get good service out of it. I am writing this on it.
My first D0S computer was a PC; 256 kB on board, Twin 5.25/360 kB floppies, composite green screen with a Star 9-pin dot matrix printer. $4700 (in 1983). But I had to have it. I lusted after Multimate, which came on two 360 kB disks in those days. One each for the program and the dictionary. Right on the leading edge of technology we were.
Of course I soon upgraded it to a massive 384 kB. I am a glutton at heart. It took a year or two before I was offered a cheap cheap $500 20 MB hard disk and I was in the big time. To celebrate my coming out I added the other 256 kB. Paradise!
Looking back over the years since those heady days I can't say that I have needed any more computing power. Sure I have changed the hard disk when the original one failed and I acquired a CGA colour monitor. After a bit of horse trading I cobbled together another XT out of the best of the bits and pieces I acquired. So what I have now does not resemble my original purchase. It even has a turbo motherboard, and 2 MB of expanded memory. I told you I am a glutton.
Fundamentally however it is still an XT and not greatly functionally different from what I had all those years and years ago.
They say that a week is a long time in politics and a year is an era in computing. What I have amounts to an antique if we are to believe the salesmen.
I should, by force of the media hype, be feeling socially inferior and functionally deprived because I don't have a you-beaut-double-back-somersault-all-the-gongs-bells-and-whistles-all-singing-all-dancing-computer-
with-a-partridge-in-a-pear-tree.
We are led to believe that anything much more than two or three years old is old old technology. It has been superseded and is probably no longer supported by the dealers.
In spite of all this I cannot see what I would do with a new computer that this one does not do. I suspect that I am not alone in this attitude.
While I was being buffeted by the winds of change and sales hype a golden thought came to me that inspired this article.
How come, if the pace of change is so frantic, that we are being pressured into buying what is now five-year-old technology. That's right, the heart and soul of these latest and greatest are five year old clunkers. The 80386 chip was released in 1987. That's five years ago. In not much more than the previous five years we've had the 8086, 8088, 80186 and the 80286.
The advent of the 386 chip moved computing from 4 bits to 32 bits in just ten years. Now half of that time has elapsed since the last major release.
The 486 chip is not a new chip. It is a 386 processor combined with a 387 co-processor on the one chip. It is the same technology that was released with the 386 chip.
The 486SX is a cynical exercise in salesmanship. It is a 486 with the co-processor disabled, which makes it a 386. Oh, so there is a small cache on the chip too but that does not make it a serious upgrade.
While we are on about cynical salesmanship, I suppose you know that the 487SX which is supposed to be a co-processor for the 486SX is actually a 486 chip with an extra pin and circuitry to turn off the 486SX chip.
All this gives the salesmen lots of numbers to toss about but we have not seen a new processor for almost half of the Computing Revolution, if we calculate it from the advent of the IBM PC in 1981, which is a common concept even if it is a bit meaningless when you take a close look.
Intel, the company that designed and manufactures most of the 80xxx chips that power most of our DOS computers, has a 586 chip in the pipeline but delivery dates are still a long way into the future.
If this 586 chip turns up in, say, another 12 months what can we predict for the arrival of the 686. About the turn of the century going on past performance. Contrary to the popular perception processor development is slowing down.
What the 386 chip did was spawn a host of peripherals, many of which were not possible before. It made more memory feasible so memory got faster which in turn required faster hard disks.
Improvements in screen quality were coming anyway as were many other improvement but the 386 was a great facilitator. It has taken this long for it all to come together.
It is not unreasonable to expect that the 586 will have a similar effect. Will it take another 5 or 6 years before the 586 is the flavour of the month.
In a way the 386 is only the second major improvement in chip design. The 286 certainly was an order of magnitude better than the 8088 the software writers ignored it, treating it supercharged 8088. There were only two significant programs written for the 80286 chip. They Windows 286 and OS/2.
The hardware manufacturers certainly did a lot with the 286 chip but not a lot of what they put into the AT architecture required a 286. Except the bus of course. A lot of what became standard features did not require that particular chip. The AT was a platform for the latest and greatest.
Now we are in a position where we have had this prodigiously powerful 386 chip. Do you
know what you can buy now for $2500 represents more computing power than you could buy for
a million dollars in 1980. That's with 1980 dollars too
And what are we doing with this super powerful chip that is already five years old. We are
using it to soup up the old AT architecture. That's right. We are putting 386 chips in Industry
Standard Architecture.
This is the board that was designed for the AT way back in 1983. The 386 is a 32-bit chip in
a board that was designed for the 16-bit 286.
There are very few Extended Industry Standard (EISA) boards around. They cost too
much and there is still not the support for them in the way peripherals or industry know-how. They and
the Micro Channel Architecture are at least capable of using the more advanced features of the 386
chip.
That's where IBM fell in a hole. They assumed that everyone would swap to the latest and greatest
and follow IBM's lead. They didn't. Everyone stayed with what they had and with what they
knew were familiar with. They wanted to get the most out of their current investment. And the
first versions of OS/2 were a bit flakey, which rubbed off on MCA, which is the architecture of
IBM PS/2, which is a 32-bit architecture.
Which brings me back to my ruminations my poor old XT. I don't need the latest any greatest. I can do all I do on this
workhorse OK if a gentleman comes in through the flywire or it dies of old age I will buy a more up to date machine. Meanwhile I will do what the rest of computerdom seems to be doing. I will plod along doing
more of the same in a well supported tried and true environment. Getting the best value for my
dollar
by ignoring all the hype.
I now derive great comfort out of the realisation that the world is being conned into buying
a five-year-old chip in a nine-year-old architecture to generally do what I am doing on my old XT.
I hope that noise I can hear is my XT laughing with me.
Reprinted from the April 1992 issue of PC Update, the magazine of Melbourne PC User Group, Australia
