The item I wrote about the Epson HX-20 the other day reminded me about one of the other items I was supposed to try to sell for that business supply company, the Epson QX-10. This beastie:
This particular example is, judging from the color of the case, an elderly one. The plastics used for computer cases rather rapidly turned an unappealing shade of dirty yellow. In its prime, though, it was a rather handsome creature, and it was both one of the most advanced, and the most useless computers I’d ever worked with.
At the time the QX came out the computer market was going through a shakeup and even more importantly, a shakeout. There were dozens of different computer makers back then, offering an astonishing variety of systems that ranged from the silly to the sublime. But at the time, IBM with its PC and the MS-DOS operating system was well on its way to becoming the standard for small business and, eventually, home computers. By the time Epson brought the QX-10 to market, its underlying hardware was already pretty much obsolete, and it’s sophisticated software and graphics weren’t enough to make up for it’s lack of horsepower.
Before IBM jumped into the market with the PC, the ‘standard’ for small business computers was the 8080 or Z80 CPU based microcomputers running the CP/M operating system. These computers were based on an 8-bit CPU and limited to 64K of RAM. Then IBM came along with it’s PC, which used the 16-bit Intel 8088 which could handle up to 640K of RAM, at around the same price as the 8-bit CP/M machines, and the rest is, as they say, history.
How did Epson hope to compete in a market that was already crowded with other 8-bit, Z80 based computers, or to compete against IBM and MS-DOS?
By coming out with a operating system of their own which was combined with a hardware package that made the QX-10 the most sophisticated system ever produced. Or so they claimed.
The QX-10 was admittedly pretty sophisticated. It had a high-resolution monochrome graphics system with up to 128K of dedicated video memory that blew away anything except dedicated CAD systems. It’s Valdocs operating system was incredibly advanced for it’s day with a built in Help system, 128 character long file names when everyone else struggled along with 8 characters. And it had 265K of RAM.
And it had what was possibly the first WYSIWYG ‘what you see is what you get’ word processor to become widely available at a (somewhat) reasonable cost. Boldface a word? It showed up in bold on your screen. Same with italics, underlining, etc. Virtually every word processor on the market at the time showed not bold face, but codes embedded in the text to turn on or off control functions, if they allowed things like bold face or italics at all.
They gave me one of these things and I had it at home for a few weeks while I learned it inside and out because I was supposed to support the thing. It was definitely sophisticated. The graphics capabilities were outstanding. It was undeniably an amazing computer when combined with the Valdocs system.
The problem was that it just didn’t work very well. Valdocs and TPM, the underlying operating system, were full of bugs. It seemed every other day I was getting updates and bug fixes. And since this is the pre-internet, that meant either dialing the company’s BBS system with a 300 baud modem and paying long distance phone bills, or waiting until they shipped me a floppy disk with the updates.
The biggest problem though was it was slow. Oh dear lord it was slow! Any kind of competent typist could easily outdistance the Valdocs word processor, getting forty, sixty characters ahead of the display update. So far ahead that you could easily overload the buffers and lose characters and words. And since we were supposed to push this as a word processing system because of the WYSIWYG display system, well, it’s pretty hard to sell a word processor that made you work slower.
The other problem was that there was no software for the Valdocs system except what was supplied by Epson. The word processor, calculator and drawing program and, I think, a rather brain dead database. There was a spreadsheet but it was so abysmally slow you could go get a cup of coffee while it was recalculating.
If you wanted to use it for actual work, that meant you had to reboot the system with the old CP/M operating system to actually do anything useful. And, of course, once you booted into CP/M, all of the fancy features Epson was pushing were lost and all you had was a generic and overpriced CP/M computer.
Then there was the competition. At the same time Epson was pushing the QX-10, the IBM-PC was becoming the standard for small business computers. There was lots of genuinely useful business software available for it. So basically there was absolutely no reason to buy the QX-10 with it’s outdated hardware, useless Valdocs system or the increasingly obsolete CP/M system.
Epson’s solution to the competition from IBM was to find someone to supply them with a plug in card that was basically an IBM-PC clone on card, while they scrambled to get the QX-16 system on the market. This ‘solution’ was literally a PC clone on a card that plugged into the computer’s internal bus, with an 8088 CPU, it’s own memory, everything. It worked, sort of. But it didn’t actually run MS-DOS, it ran PC-DOS which was an MS-DOS clone. It would run some MS-DOS based software. Sometimes. Maybe.
It also cost in the neighborhood of $1,500 if I remember right.
So you have a computer with a base price of around $2,500, already far more than comparable CP/M machines. And now you have to drop another $1,500 for a card to make it use MS-DOS software, and there’s no guarantee it will actually run the software you need…
Could it have been a great computer? I don’t think there’s any doubt that it could have. The QX was, on the surface at least, one of the most sophisticated systems to hit the market at the time. It had a lot of features that eventually became standard on later generations of computers; long file names, WYSIWYG word processor, high resolution graphics, etc.
Unfortunately, design decisions crippled it. The decision to go with the Z80 processor meant it would never have enough raw horsepower to live up to the hype. The graphics system’s hardware was woefully slow. The Valdocs system, while very nice, was bogged down by the obsolete hardware and inefficient programming techniques. Even worse, Epson never brought out any software that ran under Valdocs except that which was included with the computer. That meant that in order to run the popular business software of the day, the computer had to be rebooted into CP/M, and that turned it into nothing but a vastly overpriced, generic business computer.
Valdocs itself acquired a reputation of being buggy. I never really ran into serious problems with it except it’s woefully slow speed, but I wasn’t using the computer under actual business conditions.
There were rumors flying around that over at Rising Star, the company that made Valdocs and its underlying OS, TPM, programmers were routinely fired as soon as they finished work on their assigned modules, leaving people who were unfamiliar with the code to try to support and debug problems.
I was told that large parts of Valdocs and even TPM had been written in Forth, of all things. Forth is not exactly what I’d call user friendly. It was never designed for large projects. It was originally designed as a hardware control language used to control telescopes. I’m not saying it can’t be done, but oh brother… I’ve programmed in Forth and I wouldn’t want to use it for any kind of complex system.
Epson went on to bring out the QX-16, an interesting machine that was intended to compete head to head against the IBM PC. It had both a Z80 and 8088, and would run either Valdocs, CP/M or PC-DOS. Alas, it wasn’t very good either.
The upgraded hardware didn’t cure the system’s speed issues. The word processor was faster, but screen updates were still unacceptably slow. The spreadsheet was terrible. Reviews at the time claimed that a spreadsheet that would recalculate in just five or six seconds in MS-DOS or CP/M spreadsheets, would take minutes to recalculate under Valdocs. And while it could run some MS-DOS software, a lot of it wouldn’t run at all.
3 thoughts on “Computers that should have been great but weren’t”
My first computer was a Commodore VIC-20. I used it for RTTY on amateur radio. The “software” was a card that plugged into a parallel port, or you could use a cassette tape interface. It had I think 16 KB of memory and no internal storage. It was barely more than a dumb terminal, but it worked great fro RTTY. Wow, how times change. I think a modern microwave oven has more computing power than the VIC-20.
It’s amazing how far the computer industry has come since the 80s when I first got involved. You’re indeed right, the average microwave has more computing power than the Vic did.
The Vic-20 was remarkably useful despite its limited resources. If I remember correctly it was pretty popular with the amateur radio community for a variety of tasks. Hams often flocked to these inexpensive computers because they could be adapted for RTTY, CW keyers and other tasks. I didn’t get my amateur radio licence until relatively recently, but a lot of the projects I experimented with back then were related to ham radio. One of the very first circuits I built was an attempt to make a CW decoder for my Atari 800 back around that time. Never worked very well because the computer put out so much RFI that it rendered my receiver utterly useless – sigh. The amount of RFI those old computers put out was astonishing. I had a Ohio Scientific that I literally had to build a faraday cage for or it wiped out every AM radio and most of the televisions within several hundred feet when I turned it on.
LikeLiked by 1 person
That you for the bit of history. I used to have a customer who swore by Valdocs, so I got to use it as well.
One note: in your article, every IT’S – except one – should be ITS.
ITS processor …
The only IT’S should be “well, IT’S pretty hard to sell a word processor that made you work slower”