kludge—that was software for you. And what you learned first was often
your first love. A computer magazine even called one article “Getting
Married to WordStar.” “Can you erase your memory modules?” Mark
Robinson, a marketing man with Lexisoft, asked before he mailed me a
copy of his company’s Spellbinder word processor. He had a point. I
liked Spellbinder more than Select, much more, because it ran faster on
the Kaypro, but I still could not abide by the need to enter a new mode
to make insertions or corrections. It seemed inhibiting, this
requirement, though perhaps more in writing than in secretarial work.
Michael Canyes, a computer consultant, at the time loved Palantir. It
offered many of the better features of both WordStar and Spellbinder and
let a nonprogrammer set up the numbers pads of many computers to serve
as function keys. Canyes also appreciated The Final Word, an excellent
word processor for footnoting and other academic requirements. It was
clear. There were serious alternatives to WordStar—some, anyway—even if,
like the writer of the “Married” article, I was already well matched.
In some ways, however, I felt uncomfortable about the general direction
in which the software business seemed to be moving. Would less powerful
but well-marketed products like Select drive out the future WordStars?
_The_ WordStar had benefited from Seymour Rubinstein’s salesmanship. But
Rubinstein also had boasted a solid computer background; and unlike some
of the new people in software, he wasn’t just a businessman cashing in
on an exploding market.
Again and again a pattern was beginning to repeat itself.
Programmer-entrepreneurs would tough it out in the proverbial
garages—for some reason programmers work in garages, authors in
garrets—and score with the right software hit. Propelled by little more
than their drive and technical expertise, their companies would grow.
Then, however, other businessmen, real businessmen, would want the same
niche, and the marketing people would muscle in and set the tone. Or, as
with some new companies, they might have been setting the tone from the
very start.
This had been the classic pattern in many industries—Orville and Wilbur
Wright hadn’t exactly gone on to become major aviation magnates—but
computers were improving faster than had airplanes. And so the process
was accelerated; well-financed managers were rapidly replacing
garage-style entrepreneurs.
The stakes were higher, and striking back, software authors were even
hiring agents. The question arose, however, of the extent to which the
market would remain open to more imaginative garage people, anyway, what
with increased use of teams to develop complex programs.
“The frontier,” said Jeffrey Tarter, publisher of _Soft.letter_, an
industry publication, “is already closing.
“A couple of years ago you got to be successful primarily by competently
executing an innovative idea and by picking the right machine to produce
that execution for. VisiCalc for the Apple is the classic example here,
but lots of other programs achieved success on a smaller scale by doing
something useful for a microcomputer that came to dominate a market
niche”—or, I would add with WordStar in mind, by producing a program
valuable for its very ability to run on a number of machines.
“Now, however,” Tarter said, “the rules are changing.... Marketing
factors are becoming more important. You have to package and distribute
the product with far greater sophistication than before and be able to
spend the kind of money that sophistication requires.”
“It’s like the record business,” said Nick Vergis, then a marketing man
with Perfect Software, telling me about the new programs his company was
developing. “You get a lot more bullets ready than you actually fire.
Some will be hits, and some won’t. You hope you’ll get enough hits to
pay for the new bullets.” Research and development (R & D) was a major
part of the costs of new programs. The original WordStar had come
cheaply; Rob Barnaby had toiled night and day for a fraction of the
money that his counterparts now often demanded.
Then again, advertising and other non-R & D expenditures in the industry
had increased, too. By spring 1984 Select had sold 50,000 copies. That
should have been enough to sustain a small or medium-sized software
company, but even after a public stock offering in November 1983, Dean
still felt capital-short. “We just couldn’t spend enough on ads and
still have money to pay the sales force,” said Dean, who “saw it coming”
by February 1984. That summer he licensed Select to Intelligent Systems
to publish and promote at the retail level. Intelligent, privately held
and based in Georgia, enjoyed yearly sales of $90 million; its Quadram
subsidiary made well-advertised computer accessories. In late 1984
Select merged with Summa Software Corp., an Oregon firm that offered a
stock-market program called Winning on Wall Street, and Dean laid off
six marketing people from Select, so that his original firm had only
twenty-one workers left. The Intelligent deal went on despite the
merger.
“There are many more products available today than the market can
support, and it’s very confusing to consumers,” said an Intelligent
vice-president named Frank Marks. “The user needs to know that he’s
buying from a company that will be around in the future.” He was right.
In that sense, the large companies’ growing ascendancy in software may
have been to the benefit of the consumer. Still, old micro hands may
have gnashed their teeth when they heard that Intelligence would turn
Select into the “Quadsoft” line of software and when Marks said that the
name “leverages Quadram’s brand awareness.” Quadsoft might as well have
been a new shampoo.[25]
Footnote 25:
The account of Select’s capital shortage and the metamorphosis into
Quadsoft comes from _InfoWorld_, August 13, 1984. The merger
information is from _InfoWorld_, October 8, 1984.
To Select Information Systems’ credit, the company had garnered some
favorable reviews of new products in the months before the Quadsoft
deal. (They included Select Write, a $99 stripped-down Select.) Even so,
some of the old-time micro people correctly questioned Dean’s style of
“user friendliness.”
In a _Byte_ article, “Simplify, Simplify, Simplify,” Dean had pushed
=menu-driven= software like Select.[26] The idea seemed logical. You’d
see the choices on the computer screen rather than having to recall
combinations of keystrokes, as you would with =command-driven= systems.
Dean likened a computer menu to a restaurant menu. He lampooned the idea
of “a command-driven restaurant” in which, supposedly, you’d try “to
remember what sort of sauce the veal came with last time and whether it
was pepper or peppercorns that you liked in your green beans.”
Footnote 26:
Dean’s _Byte_ article appeared December 1983, p. 161.
Wasn’t that stretching it, however? If you ate at the same restaurant
day after day, couldn’t you order many kinds of meals without bothering
to glance at a menu? How would you feel about a waitress who _insisted_
you see a menu even if the fare never changed? She would be pushy,
wouldn’t she? All of a sudden “user friendliness” would be antiuser
pushiness to those using a program regularly. Strictly menu-driven
software should normally be just for occasional users, children, and
adults wanting to be treated like children; and many if not most adults
in the business world worked too often with words and numbers to let
Select-style programs keep them in the infant stage. Granted, exceptions
existed justifying this kind of software. A micro user who occasionally
retrieved information from a mainframe, for instance, could take
advantage of menus to guide him through complicated routines. What’s
more, even WordStar employed a menu to handle installation procedures to
match the program up with various computers and printers. But that was
different from menus that slowed you during your regular work with the
program.
Some of the so-called command-driven programs, moreover—at least
WordStar—actually did have the option of letting you keep a list of
common commands on the screen if you wanted.
Jerry Pournelle, a prominent computer columnist, growled back at the
“user friendliness” crowd, saying that their software didn’t let
ordinary people enjoy the full power of their machines. “It takes a
little work to join the micro revolution,” he said, “but the rewards can
be high.”[27] I agreed. You didn’t have to learn programming, but you
owed it to yourself and your company to make good use of the best
software for your job. A user named Alan Scharf had done exactly
that—saving his firm $200,000 a year.
Footnote 27:
Jerry Pournelle’s rebuttal to Dean and others of the “user friendly”
school appeared in _Popular Computing_, May 1984, p. 81.
Backup:
◼ IV, On the Evolution of Software (And a “Perfecter and Perfecter”
Program), page 310.
6 ❑ Three Software Stories: Motorcycles, Homes, and The $200,000-a-Year
Disk
Zzzzzz. I doze off reading lists of the right software for, say,
beekeepers or chiropractors. So sorry; I won’t cover everyone’s
programs. Besides, software is evolving. I hate Select now; I may love
its future editions. And what about WordStar? Although 2000 offers many
improvements over my beloved WordStar 3.3—and I may eventually convert
to it—I don’t like all the changes.
_So don’t think specific programs here. Think concepts._
_In fact, if possible, don’t even think computer concepts. Instead,
think_ business. _Just how could the lessons of the three people in this
chapter apply to_ yours?
Ed Boland: Accounting
Ed Boland helps run a motorcycle shop, but you’ll never find him roaring
off on a bike, dodging myopic drivers. “I’ve got a wife and three kids,”
he joked. “Scares the hell out of me.” It doesn’t matter. Boland, a
graying, middle-aged accountant, is just as useful to Clinton Cycles as
a top mechanic or a crack Suzuki salesman.
Above all, a good controller obviously knows where the money is. He
helps his employer avoid unneeded trips to the bank to borrow more.
Clinton Cycles, in fact, borrows rarely.
That’s one reason why, when U.S. motorcycle sales suffered the economic
equivalent of a head-on crash, Clinton survived. The firm even expanded
during the early 1980s. Also, years before, owner Don Smolinski had
wisely diversified. Clinton Cycles was actually Clinton Cycle and
Salvage, Inc., a miniconglomerate with three motorcycle stores and
dealings in scrap metal. Sales exceeded several million a year.
But size and complexity weren’t pure joy. Boland’s accounting software
just couldn’t handle a holding company and five subsidiaries, not
without costly customizing, and nearly a year after computerization, his
feelings were still mixed.
“I don’t regret the decision,” Boland said. But at the time he had spent
thousands of dollars more than planned, and he told me, “There’s been a
few times I would have thrown this computer out the door in a second.”
His horror story—even if it eventually turns into a success tale, as it
probably will—is instructive. He’s good. And he bought his software from
a good store. Clinton Computers, in the same Maryland suburb, won top
ratings in a consumer guide in the Washington, D.C., area. A software
executive, moreover, an outspoken one critical of some other D.C.
dealers, praised Clinton.
So a bothersome issue arises. If a $3,800 accounting program doesn’t
work out easily with good people selling and using it, what happens
among mediocre people?
And there is a second point here. Don’t expect even the best computer
store to serve as your private consultant, intimately familiar with your
business; not, at least, unless you’re buying a big mini or mainframe
from the likes of IBM.
A third point, too, comes through. Don’t computerize a
multi-million-dollar business without a consultant—at least not if
you’re running an extra-hairy program like an accounting one—unless
you’re willing to talk and think computerese or at least take a training
course. Accounting software, anyway, just isn’t simple and reliable
enough yet. Any idiot can flick on a computer and stuff a floppy disk
into a drive or crank up a hard disk; most of the time his hardware will
work fine, but even the brightest of businessmen may need a consultant
to unravel the mysteries of many programs—even “canned,” off-the-shelf
accounting packages like the one that bedeviled Ed Boland.
A blunt, tough-talking man, Boland himself is no moron. He holds a
bachelors degree in accounting, and he’s worked for employers ranging
from hospitals to restaurant chains, including one where he caught a
chef on the take from a supplier.
“Accounting is accounting,” he said. “I don’t care if it’s food or
motorcycles. Life exists the same way.” Moreover, life is a series of
ratios, of intertwining figures. That’s how Boland caught the chef. He
knew that food costs were normally 20 percent of the company’s expenses;
two months after the chef’s hiring, they were 23 percent, creeping up
toward 27 percent, finally. The incident perhaps colored his attitude
toward computerization. He would computerize, but not overdo it, not
lose himself in tangential statistics, not hire extra clerks to key in
data.
“We carry more than seventy thousand new parts _numbers_ alone,” he
said. “Seventy thousand different kinds of screws, bolts, nuts, fenders,
rims.” But if he couldn’t track every single bolt, he at least could
work with major numbers like sales figures for various categories.
Normally, for instance, new bike sales and accessories sales went up and
down together, and if one of the two had fallen off alone, you could bet
that Boland would demand a reason. If nothing else, computers, by making
past information available more quickly, could help him sniff out
trouble.
Computerizing, he saw a micro as one way to avoid greedy consultants and
their expensive recommendations.
“I’ve seen too many instances,” said Boland, “where businesses paid out
$50,000, $60,000 to a consultant who spent six or seven months analyzing
computer requirements. Then he’d recommend $50,000 or $150,000 worth of
computer equipment.” Boland wanted the benefits of the bigger machines
without the costs. His goal was to reach the point that a somewhat
larger competitor in the D.C. area already had; the rival owned an IBM
34 system with a six-digit price.
“He used to work for IBM fifteen or twenty years ago, before he went
into the motorcycle business,” Boland said. “He now has just about every
one of his salespeople on commission, and he knows exactly what they
sell every day. He can tell you where each motorcycle, each piece of
inventory, is.
“He can tell you whether it’s in a truck, what color it is, where it is,
what stage of preparation it’s in for delivery. He puts his purchases in
the computer with an inventory program. The moment a salesman sells a
cycle, the register itself deducts it from inventory. It’s a
=point-of-sales= system. So theoretically, at the end of the day he
knows exactly what he started the day with, what he sold, what he added
to it, and what the bottom-line figure is. We’d love to get to that
point.
“The likelihood of this happening with us, though, is very low. I don’t
know if all this detail is necessary. And our business is seasonal, and
I don’t want too much cash tied up in the system. And what’s the sense
of keeping instant track of every nut and bolt sold? But we’d love to
get the major items in on computer. The motorcycles, for instance. They
can easily sell for up to $6,000 apiece, so it’s worth it.
“It’s a question of software. This competitor has spent more than
$100,000 to develop specific software to get what he wants. Now if
you’ve got $100,000 to throw away and want to hire a programmer full
time at, say, 0,000 a year, that’s fine. Most companies like us can’t
afford that.”
So Boland bought several programs off the shelf from Clinton Computers,
including Accounting Plus, which he began trying to use as an electronic
general ledger. He ran it on a North Star computer with a 5-megabyte
hard disk—a memory device capable of stashing away the equivalent of
maybe 2,500 typed pages of information. The North Star system was just a
way for him to get his feet wet in computing with a general-ledger
program. Boland, though, would switch computers even sooner than
expected. The system’s memory space wouldn’t suffice for the records of
his miniconglomerate, and it wasn’t easily expandable. And he would have
trouble getting the machine repaired as quickly as he needed. Clinton
Computers, however good its service department, just could not respond
fast enough to suit _his_ requirements.
And the software seemed just as major a disappointment, what with the
need for complicated customization.
“The biggest problem I ran into at the store,” he said of the hardware
and software, “was ‘We’re going to sell you a system that we think will
do the job, but not necessarily so.’”
Both the buyer and the seller, in this case, apparently lost the
software gamble; they apparently came out on the wrong side of
“necessarily.” Clinton Computers did not win, because it _isn’t_ a
fly-by-night operation. Boland’s name had come to me from none other
than Sue Grothoff, who’d sold him the software. She had enough
confidence in herself and her store to offer Boland as the source of a
“candid” story, and my respect remains, especially after having talked
to another business customer, raving happy about her ability to
understand his company’s software needs.
Boland emphasized that Clinton wasn’t out to cheat him. “Sue is a very
nice person.” And he liked the software expert there, too. “He knows how
to change software, but you have one person trying to meet the needs of
too many customers. He’s spread too thin. Clinton never gave him enough
time to get the program running at my business. And we never could seem
to get together to get the twelve hours of training they’ve promised me.
To be frank, I think we spent the twelve hours trying to correct the
problem I had rather than on training.”
Giving Clinton Computer’s side, Grothoff said, “Because of changes that
Ed wanted in the software, it took longer to adapt the program to meet
his needs. And that used up the twelve hours.” Evidently, Boland and
Grothoff stayed on good terms, because she later sold him a new Kaypro
portable for home use.
But with his big computer system at work, Boland—for one reason or
another—seems to have suffered more than his share of woes.
“Accounting Plus,” said Boland of the software he bought there, “is a
very good program, but it was designed for one single company to be
subdivided into departments. You could have six departments, everything
from sales to used parts, and it will work beautifully. But it wouldn’t
for what I wanted. It would not treat a motorcycle store, for instance,
as a subdivision of a holding company. I could not pull off separate
profit and loss statements or separate balance sheets.”
Clinton Computers, to its credit, referred Boland to a consultant
familiar with the =source code= of Accounting Plus—a part of the program
that would enable it to be customized. The consultant took in Boland’s
North Star computer system as a trade-in. He sold him a Delta
microcomputer and TeleVideo terminal in return, along with a hard-disk
system upgradable to 70 megabytes. With the software customization
included, the cost came to $20,000, minus the $12,000 trade-in. The
consultant, like Clinton Computers, didn’t know as much about accounting
as Boland had hoped. But through sheer tenacity Boland at least got
himself a halfway usable program.
“I had the software changed,” he said, “so that instead of one company
with departments in it, it now reads a holding company with subsidiary
companies. The software was changed to print out separate balance sheets
and separate income statements for me.
“On the income-statement side this machine will ask me if I want a
consolidated or individual by-page income statement. It will print out a
holding company and then, in successive order, on each page, the income
statement for each company. When I go to the balance sheet, it will ask
me that same question.
“If I tell it I want individuals and a consolidated by page, it won’t do
it. I must ask it individually each time for company one, then company
two, right down the line. I have to give the computer individual
instructions to do that.
“And that was the best my consultant could do to change the software. It
got the job done, but not all the way.
“I’ve since found out that for $400, not ,800, I could have bought an
accounting package that would do exactly the same thing I wanted.”
So why didn’t he buy it?
“Why spend another $400 to delve into it further and then find out _it_
isn’t what you wanted?” Boland asked. “I think it’s stupid to spend
money at this point. I have a system I know I can make work within
certain limits. I’m still in the learning phase, and I’ll probably still
be there for another year, for all I know.” I thought he was too
patient. If his software system was a dog—perhaps not for everyone but
for him—he should change it. Four hundred dollars was a pittance
compared to Boland’s total investment of $20,000.
Ideally, too, as he said himself, he would have gone to a consultant in
the first place, not just any consultant but one familiar with
accounting. And yet it’s understandable why he acted as he did. Why
repeat the mistakes he’d seen at other companies? Why not frugality?
Better to be out $20,000 than $50,000 or $100,000 after paying for a
mini, a consultant, and the other trimmings. His computer wasn’t making
or saving buckets of money for Clinton Cycle and Salvage as of early
1983; but it might in the future as he got his software under control
and could, for instance, easily put salesmen on commission.
He told me, moreover, that he would soon stop farming out the company
payroll—up to seventy people—to a computer firm. He also was mastering
an electronic spreadsheet. He did not plan an accounts receivable
program, because Clinton Cycle collected quickly through checks and
credit cards, but an accounts payable one was in the offing.
“It’s going to give me the companies’ names, the payment terms, the
volume I deal with them each month,” Boland said. “It will flag when the
bills are due, and it will also generate checks.”
He was also going to crank up a mailing-list program, using a list of
all buyers of new bikes within the last three years—a sensible project,
considering the market for accessories and repeat sales. Obviously,
here, as in other applications, the more information built up on his
hard disk, the more his machines would justify the $20,000 investment.
The benefits of computerization for Edward Boland might not be
immediately dramatic, but with his stick-to-itiveness, sooner or later
they’d almost surely come. Boland’s battles with computers were to
continue. When I caught up with him in spring 1984, he said he’d bought
yet another computer system.
Charlie Bowie: Data Base and Spreadsheet
Remember the customer who was raving happy about his software—also
bought from Sue Grothoff at Clinton Computers?
That’s Charlie Bowie, who, when we first talked, was a vice-president of
Washington Homes and manager of its Southern Division. He told me he
expected two Zenith micros to save his company perhaps as much as
$50,000 a year, maybe more.
In just one month, in fact, his $5,000 computer had already saved the
equivalent of thousands of dollars in executive man-hours.
Even before he and I met, his assistant, behind his back, was singing
praises of her boss and their new machine. “I wanted to learn word
processing,” said Julie Grimes, a young, blond woman who could have been
the secretary in the IBM commercials saying the advertised product was
“a piece of cake” to master.
Gazing at the Zenith—which she and Bowie use for data-base work,
spreadsheets, and some word processing—she said: “It really comes in
handy.... We were both worried. Charlie told me, ‘Don’t worry. We’ll
teach each other.’”