Is the QX—16 with Valdocs
a "fantastic machine" or an
"embarrassing failure"?
COVER STORY by Jonathan Sacks
There was precious little excitement this
past October when Epson America
called a press conference to announce its
second-generation personal computer,
the QX—16. In too many ways it seemed
like deja vu. While the Epson name has
come to represent the standard in dot-
matrix printers (with an estimated 40
percent of that market worldwide), the
company’s one desktop computer to
date—the QX-10-can only be classified
a failure.
The new Epson machine is essentially
an updated QX-10 with IBM compatibil-
ity thrown in for good measure. It bears
so many harrowing similarities to Ep-
son’s earlier machine—including a de-
pendence on a troubled software pack-
age called Valdocs-—that one industry
insider asked an Epson official whether
the whole thing wasn’t a bad joke.

61

To understand that kind of reaction, you have to know
a little bit about the rise and fall of the QX-10. That first-
generation Epson, released with great fanfare and a
huge advertising effort in late 1983, was elegantly
designed, with a spectacular keyboard and a crystal-
clear bit-mapped screen. But it was built around an 8-bit
Z8O microprocessor at a time when the 16-bit 8088-based
IBM PC was fast becoming the standard. Even worse,
the QX-10 had a $2995 price tag—hundreds of dollars
more than other 8-bit machines.
Epson’s only hope for the success of the QX-10 lay in
a revolutionary software package called Valdocs, which
came bundled with the computer. Valdocs featured a
word processor, a calculator, a "draw" function (that
created bar; pie, and line graphs), an electronic mail func-
tion with an electronic address book, and a computerized
date book.
It also had a user interface that was meant to change
the way people worked with computers. Its designer,
a visionary named Chris Rutkowski, wanted the com-
puter and the software to be so simple that a neophyte
could operate it without opening a manual. He designed
the QX-10 keyboard so that, when used with Valdocs,
all computer functions could be activated by hitting a
single key.
And the keys were labeled in English. A key labeled
Menu gave you a menu, a key labeled Draw put you into
the draw program. 'Ib store a document, you hit a key
labeled Store. The text editor was "what you see is what
you get." You could boldface or italicize (on-screen and
in print) by hitting a key labeled Bold or Italic. You got
extensive on-line help for each application just by hit-
ting a key marked Help.
Valdocs was brilliant in its simplicity, and immediate-
ly following its announcement, the reviewers went wild.
But soon problems began to crop up. Entire documents
would disappear in Valdocs for no apparent reason. The
electronic mail function usually didn’t work. And the
whole thing was painfully slow—it took a half minute
to delete a block of text and a full minute to delete a file.
Epson and Rising Star, the American software com-
pany contracted to write Valdocs, promised fixes and
upgrades. Then they missed their own deadlines. Within
months after its introduction, the reputation of the
QX-10 was forever stained by the debacle of Valdocs.
Even Epson’s efforts to sell it with a free printer and
additional software failed to bolster sales. A year later,
the machine was widely discounted to $1595, and QX-10
unit sales still accounted for less than one percent of
all personal computer sales.
A New QX
And so, on that October day in Torrance, Epson had
a grim history to overcome. High-ranking company of-
ficials flew in from Japan to join American executives
in hours of windy speeches that reaffirmed Epson’s com-
mitment to carving a niche in the American computer
market.
But when at last they introduced their new machine,
it seemed that in ignoring the lessons of history, these

62L

men were doomed to repeat their embarrassing failures
Like its predecessor, the QX-16 uses a Z80 micropro-
cessor to run CP/M and a new version of Valdocs, called
version 2.0, which is slated for release by the time this I
article is in print. In its sluggish way, Epson has also
finally given its computer IBM compatibility by in-
cluding an 8088 coprocessor and a special graphics card
that allows it to emulate the IBM PC. Epson claims that
the QX-16 is capable of running 90 percent of the pro-
grams available for the IBM.
The machine comes with two 5 1/4-inch slimline disk
drives, which have 360K bytes of storage each in the
8088 mode and 768K bytes of storage (quad density) in
the Z80 mode. (A hard disk is available from Comrex,
a subsidiary of Epson’s parent company, Suwa Seikosha,
a manufacturing unit within the Seiko Group.) Also in-
cluded are a Centronics-compatible parallel port and an
RS-232C serial port. The QX-16’s standard 256K bytes
of RAM can be expanded to 512K.
Although the system’s three expansion slots will not
accept IBM cards, Epson says it will release its own ex-
pansion cards in the future.
The standard monochrome monitor is driven by 128K
bytes of dedicated video RAM and an NEC 7220 graph-
ics chip. The 12-inch bit-mapped screen has a dot reso-
lution of 640 by 400 in Valdocs and 640 by 320 or 320
by 200 in MS-DOS. Epson says a color monitor and col-
or graphics card will be available in the future.
The QX-16 keyboard is a Selectric-like layout with 18
special control keys across the top. The 10-key number
pad is programmable via the "Defkey" function, which
is part of Valdocs 2.0.
At the time this article went to press, the company
had not settled on a final price for the QX-16. It prom-
ised a price of "under $3000," and the final price was
rumored to be around $2500.
The New Valdocs
In a marketplace that has clearly defined its most im-
portant ratio as price/performance, the QX-16 seemed
destined to follow in the footsteps of the QX-10. Why,
after all, would anyone want to buy an IBM compatible
for about the same price they could buy an IBM PC?
"The point is that we aren’t positioning this as another
IBM compatible} ’ responded Scott Edwards, marketing
services manager of Epson’s Company Products Divi-
sion. "We don’t think that’s its niche. It would never sell
as a compatible at this price."
So how exactly does Epson plan to market the QX-16?
"We’re depending on Valdocs," said Edwards "Valdocs
is what we have that makes us special. And the latest
version is great, just great. It will make the QX-16 a
fantastic machine at a fantastic price."
A sadly familiar refrain, except this wasn’t the old
Valdocs that Edwards was speaking of, the Valdocs that
had come to be known as oh so buggy and oh so slow.
No, he said, the new Valdocs, Valdocs 2.0, was every-
thing that Valdocs should have been the first time
around. (See ‘An Early Look at Valdocs 2.0," page 64.)
This version of Valdocs, Edwards said, has been speeded

62R

up, cleaned up, and polished. This is a Valdocs to make
Epson proud.
Epson put on a short demonstration of Valdocs 2.0 that
day in Torrance. It included some glitzy draw and paint
programs akin to what was available for the Macintosh
and the IBM PC. The program now includes a spread-
sheet and an improved mail function, which one Rising
Star official claimed was as good as any communications
package on the market.
"What we tried to do with the first releases of Valdocs
was establish that it was possible to create easy-to-use
integrated software," said Gordon Mustain, director of
marketing for Rising Star. "What we did with version
2.0 was try to make it right."
Valdocs’ Big Daddy
When you talk to people who know about Valdocs you
hear one name again and again. Depending upon whom
you ask, Chris Rutkowski is either a genius or a fool.
For better or worse he is the man who created Valdocs
and the man who knows it best. To learn more and to
get a firsthand demonstration of Valdocs 2.0, I visited
Rutkowski at his Carmel, California, home.
Rutkowski was the first American employee of Ep-
son, and he claims credit for the marketing savvy that
put Epson printers on top in the United States That,
he says, won him the respect of the Japanese company’s
hierarchy, and when Epson decided to build a computer
for the American market, Rutkowski’s offer of help was
eagerly accepted. ‘
Rutkowski’s business card identifies him as a futurist,
and from the outset of Epson’s computer project, he took
the long view. Plenty of computers were already out
there that performed adequately, Rutkowski figured.
What was missing, what was necessary before people
would make computers part of their lives, was a way
to make these new machines easy to use. And so, the
concept of Valdocs and HASCI was born.
HASCI, an acronym Rutkowski coined, stands for
human applications standard for computer interface, and
as the name implies, Rutkowski thinks that someday
every computer will have to meet his standard. VVhat
Rutkowski intended to create from the outset was an
easy way for humans to relate to computers; an inter-
face so simple that even a child, or "a little old lady in
Dubuque" as Rutkowski would say, could sit down at
a computer for the first time and make the thing work.
What he now claims, but didn’t really make clear
earlier on, is that the process would take some time and
would be done in three phases. Phase One—the one we
have seen—was the first series of Valdocs software.
Rutkowski says it wasn’t nearly as bad as some people
thought. Phase Two—the version of Valdocs about to be
released—will perfect each module. The final phase,
Valdocs 3.0, will be a fully integrated, fully perfected
program—the complete computer program, so to speak.
The process is evolutionary, and improvements are
ongoing, Rutkowski says. What burns him is that the
whole thing has been so misunderstood and maligned.
"A lot of the negatives have been grossly exaggerated?

63L

he said. "You know, it’s    
easier to criticize than to 
produce, and most of the if 
Valdocs have never made    
any contribution to our 
Society.
"The fact is that the 
criticism came mostly from computer freaks, and
Valdocs was never meant to make the computer freaks
happy. We find that there are certain types of users-
those to whom ease of use was important—who loved
Valdocs, despite what the critics were saying.
"On some levels, like ease of use, I would put Valdocs
up against anything out there. While everyone is wor-
rying about making people computer literate, we are
really trying to make a people—literate computer. Some-
day, all computers will have to be that way."
And what of the specific criticisms, the fact that
Valdocs was slow and infested with programming bugs?
"We got the bugs taken care of pretty quickly, and I
agree that the first release of Valdocs should never have
gone out to the public," Rutkowski said. "Some people
said that a program as complex, as huge as Valdocs,
couldn’t work on a Z80. Well, I have said all along that
it isn’t the size of the microprocessor that counts, it’s
the elegance of the code. Valdocs was slower than some
other programs and that was because the code had to
be refined. It is a constant process. We have made 2.0
significantly faster than previous versions."
The Evolution
The development of 2.0 began with a meeting at the
Saddleback Inn in Orange County, California. There,
Rutkowski spent four days outlining what he wanted
Valdocs to be. The thesis was a simple one. Rutkowski

63R

wanted each module of Valdocs—each application—to be
competitive with anything in its field. In other words,
the Valdocs text editor had to be as good as Wordstar
or Microsoft Word; the mail program had to be as good
as Crosstalk.
Just as he had done for the previous version of Valdocs,
Rutkowski wrote a "manual" for the nonexistent Val-
docs 2.0, outlining what it would include and how each
application would work. In addition to an improved
"what you see is what you get" text editor, there would
be an improved mail program, a spreadsheet, a card-
file-type database, business graphics, and a date book.
And, two other programs—pieces of the Valdocs en-
vironment—would be offered at additional cost. One
would be a draw program, something sophisticated
enough so that it could be used for computer—aided
design (CAD) applications. The other would be a paint
program that would take advantage of full-color graphics.
Rising Star decided that, like earlier versions, the new
Valdocs would be programmed partly in assembly
language, but mostly in a specially designed version of
FORTH that the company called R-FORTH. This

64L

despite widespread criticism that part of Valdocs’ prob-
lem was its programming language. FORTH isn’t gen-
erally considered slow, but critics of Valdocs have sug-
gested that the sheer bulk dictates that it should be en-
tirely in the more efficient assembly language.
Rising Star hired several programmers who worked
on customizing the FORTH language for the task at
hand—a reverse process, really, of the customization that
was done during the creation of earlier Valdocs versions
FORTH is a language designed to be modified, and
when the programming began on Valdocs 2.0, the kernel
of R—FORTH had swelled from 16K bytes to 22K. It had
to be streamlined. By the time that streamlining was
done, the kernel was back down to 10K bytes.
Other programmers—some 40 in all—worked in groups
to rewrite and streamline each module of Valdocs: the
data-indexing system, the editor, the spreadsheet, and
the graphics driver. They did all their work at home.
Rutkowski, the futurist, believes that people are most
productive in an "electronic cottage" environment. Also,
he says, he wanted the best programmers he could find,
and he knew he couldn’t convince many of them to move

64R

to Southern California. Each programmer was assigned
his piece and he worked until it was perfected. Notes,
comments, and code were sent to several Rising Star
on-line bulletin boards for peer review. All code was also
sent to a computer bulletin board in New Jersey where
Rising Star Executive Vice President Roger Amidon—
who coordinated the entire effort—reviewed it, made
certain it worked with other modules, and ordered revi-
sions or rewrites
The concept remained the same throughout—to create
a state-of-the-art integrated software package that
anybody could use. In its final version, Valdocs 2.0 took
over 650K bytes of disk storage. The system contained
68K bytes of help files and more than 300 menus. (Not
even Rutkowski could establish exactly how many
menus there really are.)
By the time it reached final testing, Valdocs 2.0 was,
in Rutkowski’s opinion, a better program than any other
he had ever seen. It was better than anything on the
Macintosh, which had been introduced during the
development of 2.0, because it didn’t depend entirely on
the mouse, which was, according to Rutkowski, a fine

65L

input device for limited circumstances, but not a replace-
ment for the keyboard. (Anyway, Rutkowski’s new Val-
docs uses a mouse for what mice do best——paint and
draw programs.)
Valdocs was better than Lotus Development’s 1-2-3,
not because it was more powerful, but because it was
usable by people who weren’t spreadsheet geniuses "I
have a spreadsheet that people who have never used a
spreadsheet can sit down and use,’ Rutkowski said.
It was better than Wordstar because it was a what-
you-see-is-what-you-get text editor. If you wanted some-
thing in boldface, you pushed the Bold key and it showed
boldface on the screen. It was, in Rutkowski’s eyes, state
of the art. Period.
"We have set the standard that all others will have
to match,’ Rutkowski said. "Especially in ease of use.
I don’t think there’s anything out there that’s better.
We’re talking magic."
And yet, Rutkowski admitted, Valdocs 2.0 wasn’t
perfect. While the text editor did give you on-screen
bold, italics, three sizes of type, and superscript and
CONTINUED ON PAGE 167

65R

CONTINUED FROM PAGE 65
subscript, it was still slow.
The business graphics
Were good, but the graphics,
although they could be
scaled from postage—stamp
size to full screen, could
not be moved into the text
editor.
The spreadsheet figures in Valdocs 2.0 could be
translated into graphs but again they couldn’t be moved
into the text editor or the card file. "That’s the next
step," Rutkowski said. "It’s evolutionary."
The Unveiling
Like other versions, Valdocs 2.0 was late in arriving-
months later than it was supposed to be. Six weeks
before its proposed late-1984 release date, various bugs
were still floating around within Valdocs 2.0. The reason
for this, Rutkowski said, is that a version would be
stabilized and readied for release, then Epson would
delay release, and then Rising Star would add or change
the program again, creating new bugs.
That didn’t explain why six weeks before the proposed
release date the Valdocs manual was still in pieces. It
I didn’t explain why at that late date Epson had yet to
approve 2.0 for release at all.
Even assuming that the Epson QX-16 and Valdocs 2.0
make it onto the market in early 1985, it seems unlikely
they will have much impact on the computer world as
we know it today. Rutkowski’s Valdocs, while conceptu-
ally strong, is still in development—imperfect, if you will.
For reasons only he knows, Valdocs 2.0 doesn’t take
advantage of the QX-16’s 16-bit coprocessor. Despite
Rutkowski’s insistence that the Z80 can handle the
enormity of Valdocs, it seems that the processor is
straining. Valdocs 2.0 continues to suffer from speed

167L

problems, especially when making block moves in the
text editor. Because so many things go on in back-
ground—for example each time a document is stored it
is automatically indexed by date—functions such as
saving a file take a very long time.
That Rutkowski agrees work still needs to be done
on Valdocs is both reassuring and distressing. It is re-
assuring because it reaffirms what anybody who has
ever met him suspects—he’s a bright guy, worthy of the
task he has undertaken. It is distressing because Rising
Star has a history of giving us too little too late.
Because in reality, as charming as Rutkowski’s notions
are, HASCI has not become a standard, and it doesn’t
look like it ever will. One can’t help but wonder whether
Apple, with its Macintosh, won’t whisk past Rising Star
Certainly, in the realm of integration, companies such
as Ashton-Tate and Lotus Development have already left
Valdocs in their dust.
Once again, there is a lot to love about Valdocs.
Rutkowski is especially banking on his draw and paint
programs, which allow you to zoom in on the screen and
create images pixel by pixel. In demonstation these pro-
grams are wonderful, and Rutkowski thinks they are
flashy enough to sell computers.
Perhaps Rutkowski is right, and thousands—even
millions—of people out there are looking for a computer
that is easy to use, and perhaps they are willing to trade
some of the annoyances of learning to use other com-
puter programs for some of the annoyances of Valdocs.
Epson hopes so—but the company is hedging its bets
anyway. At the unveiling of the QX-16, Epson announced
that it is also starting its own software company to
develop products for the QX-16. (Rutkowski and Epson
both claim the new company will not compete with
Rising Star.)
While Rutkowski says he hopes to have a long and
lucrative relationship with Epson, he claims to be un-
concerned about whether the public accepts the QX-16
or Valdocs 2.0. The way he sees it, sooner or later the
world will recognize the value of what he is doing.
Rutkowski, remember, is a futurist. He takes the long
view. "I’m already past 2.0, and the design work for 3.0
is already complete," he says. He’s ready for other proj-
ects, and soon, he says, he will be done with computers
altogether.
Rutkowski talks, you see, not in terms of selling com-
puters, but in terms of setting historical precedents, of
bettering the lot of mankind, of being remembered long
after names like Macintosh and IBM PC are forgotten.
"I’m here to research the leading edge of technology,"
Rutkowski says. "I’m trying to create technology by
which mankind can improve existence on this planet."
It is an awesomely long way from Valdocs 2.0 to such
a heady dream. But the longest journey begins with a
single step, they say. And anyway it might explain why
some of what Rutkowski does is so mystifying to the
rest of us. After all, in his league what does it matter
if you have full integration this month or next?
Come to think of it, in his league what does it matter
whether you ever sell a computer? 

167R