Ray Noorda didn't start Novell Inc. from scratch in 1983. He bought a company: Novell Data Systems Inc. Novell Data Systems was the hopes and dreams of other people. It had been founded in 1980; grew to 120 people in 1981, but by the end 1982 collapsed to 15 people. Things were so bad in late 1982 that the products that the company built were being returned faster than they were being sold.
But those 15 people left at the company had a dream... a dream for a revolutionary product that would save the company, and Ray found that dream so exciting and real that he was willing to invest his time and money in this otherwise losing proposition to make this dream happen. What was it that Ray bought?
Those 15 people, and Ray, were the founders of the Novell Inc. we know today, and the dream was a local area network product that would become NetWare. The founders included Harry Armstrong, Craig Burton, Judith Clarke and the Superset group -- the programmers who made NetWare.
These people made Novell and NetWare. To understand them you need to understand the Novell that they watched grow and collapse. They learned a lot watching that happen. And between what they learned and the experience Ray Noorda brought, they developed a winning product and a company that made an industry. To understand Novell Inc. you need to understand the stage that it was built upon: Novell Data Systems.
The Novell Data Systems story is also important as a contrast: the goals of the companies were almost identical, but the results of the implementations were dramatically different.
In the summer of 1980 two seasoned computer industry executives got together to start a company. Like all good high-tech entrepreneurs they were starting a company that they intended would shake the world. Unlike many, their company actually did -- but hardly in the straight-forward manner that they had envisioned.
The two men who started Novell were Jack Davis and George Canova.
These men brought with them dreams and expectations that they had been building for several years. Their dreams, experiences and expectations meshed with the reality of those early 80's years to make Novell happen the way it did. So, to understand the Novell Data Systems that shaped the founders of Novell Inc., we need to go back one step further and look at the expecations these founders brought with them.
Jack Davis was born in Utah. He graduated from BYU in 1961 with History and French degrees. He first found work in the growing minicomputer industry at NCR. It was the start of a long career pursuing the Muse of Technology. He moved to Phoenix were he started working for GE's disk drive division. Later in the sixities his career took him to southern California where he joined up with Calcomp, a maker of plotters and other minicomputer peripherals, heading up international sales. From CalComp he moved to General Automation and headed international sales there.
These were the hotbed days for minicomputers and southern California contributed no less to the industry's development than Massachusetts (Route 128) or the northern California Bay Area (Silicon Valley... Silicon Gulch to wags in Massachusetts).
Jack moved through wide circles in the southern California industry in the sixties and seventies. This wide circle of acquaintances provided him with the manpower he needed to start Novell.
In the seventies Jack moved to Utah and became director of International Sales for Beehive International. Beehive at that time was one of the biggest independent manufacturers of CRT terminals. He sharpened his skills at selling peripherals and became experienced at dealing with the specialty terminal marketplaces that required something a little different and special.
The specialty market had benefits and weaknesses: the benefits were high margins for the products and fewer competitors. The weaknesses were limited market size and higher development costs.
The constraints of these specialty markets caused controversy within Beehive. One engineer I talked with years after Jack left felt that Jack was a good salesman, but he was always bring back orders for, "terminals with purple reset buttons, or some such, when he should have been bringing back orders for the stuff we made already."
Jack liked to start companies and projects; Novell was far from his first. In California he started a company to sell a protocol converter (a box of electronics that converts one communications protocol into another). This endeavor was still-born when the company he chose to manufacture the box got wrapped up in legal battles and couldn't produce.
While no product came out of this project, it introduced him to Victor Verpele, a southern California hardware designer who had done some work for an east-coast outfit named Safeguard. Victor introduced Jack to Adolf "Dolf" Pere and Pete Musser -- owners of Safeguard -- and Jack did some consulting for them. It was one of many contacts Jack made in the sixties and seventies that would be cashed in on in the eighties.
Jack's startup just before Novell was Terminal Specialties Inc. TSI was a terminal distributing company: The goal of TSI was to link terminal makers with their customers. In the late seventies many companies from the East Asian "Four Tigers" countries (Hong Kong, Taiwan, Korea and Singapore) were introducing new lower cost CRT terminals into the US. This was part of the reason the price for "dumb" ASCII terminals declined from $1500 apiece in 1978 to $900 apiece in 1982. These East Asian companies knew a lot about manufacturing but very little about marketing in the US. Jack's TSI was providing the service of getting them in touch with their market.
And Jack was effective. He added value for these companies. One of the big companies TSI serviced was Televideo. Jack negotiated a master distributor agreement with Televideo then proceeded to introduce them to big name distributors he had worked with previously, such as David Jamison Carlye out of the LA area.
Jack also worked with Televideo on improving the product's features. He helped them with big things as well as the little things that would make the product more compatible with US market tastes. For instance, he recommended changing the case color to make it more harmonious with other computer equipment.
Jack was effective, but he liked to play fast and lose and sometimes this caught up with him. Televideo was one of those cases. Televideo prospered, but in 1979 they underwent a management shakeup and ceased honoring the master distributor agreement they had with TSI. The agreement had never been formalized with a contract, so TSI was out in the cold -- left only with yet another "distributor dumped upon" war story to show for its efforts.
The Televideo story is important because it was one of the things which convinced Jack that his fortune was going to have to be made somewhere besides distributing and, like many other players in the computer field he planned his migration to be up the marketing chain. Customers dream of becoming retailers. Retailers dream of becoming distributors and wholesalers. Distributors and wholesalers dream of becoming manufacturers. Jack was going into manufacturing next.
Televideo was one of TSI's big customers. One of TSI's smaller was Dobbs and Woodbury -- a two-man terminal designing company based in Salt Lake City. Dobbs and Woodbury had designed a cutting-edge terminal for the Sperry/Univac marketplace. The terminal was good, but they were having trouble marketing it, which meant poor sales. As a result they were out of money.
Jack wanted to market their terminal, but Dobbs and Woodbury needed money right away and they were about to sell the rights to another company that was unlikely to deal with TSI, so Jack offered to buy the rights for the same price as the other company. Woodbury said yes, Dobbs said no. There were other convolutions, but the net effect was that Dobbs and Woodbury split. Dobbs retained control and Jack didn't get his terminal rights. What he got instead was a mutual committment from Rusty Woodbury to join him in his new enterprise when it started up.
1980 was not a good year for TSI. Jack discovered that in addition to losing Televideo as a revenue source, and losing rights to this specialty terminal, he had gained an unexpected expense: the company accountant had been dipping into the till. So in 1980 TSI was fighting for it's existence and in need of cutting expense. Jack and his partner agreed that the partner would continue TSI while Jack sought his fortune elsewhere. Jack needed a new organization and that organization would be Novell. It was time to assemble the team.
By 1980 Jack had informally lined up many people who were ready to help him start this new company, but he needed that which all new companies find hard to come by: money. He needed to assemble the financing.
Jack called on George Canova because he had seen his handiwork at Calcomp. Calcomp was a minicomputer peripherals maker that Jack had worked at a few years previously. He knew George had started Century Data, a prosperous disk drive maker, and that George had made a lot of money when Century had been merged into Calcomp. The merger agreement gave George some stock options which were based on Century's performance after the merger. Century did so well that George ended up Calcomp's largest stockholder.
In 1980 George had just finished an even bigger task. For the last three years as president and then chairman of Calcomp he had been turning the company around.
In the mid seventies Calcomp experienced losses. In 1977 George replaced its founder as president and chairman and worked hard to stem the "red tide". He reduced the debt load by selling off divisions (One of the hardest cuts was selling off the memory products group that had been the Century Data that he had founded.) He also worked at controlling costs in other ways. He was the first to implement a four-day work week in a major California electronics company (1200 employees).
In 1979 he had turned the corner and he was thinking about future growth in a way that was paralleling Jack Davis's thinking. In an article in Business Week, Dec 4, 1978 he talks not only about his turnaround efforts but his vision for Calcomp's future.
"Canova's next move will be into an area that he considers full of growth potential: supplying a line of peripherals for minicomputer users, who can save substantially by assembling their own minisystem ala carte.... Canova predicts this miniperipherals business venture will produce sales of tens of millions a few years out."
George's vision was controversial within CalComp, and instead reselling peripherals the company devoted resources to building their presence in the CAD/CAM marketplace which was a more logical extension of their already established plotter business.
In early 1979 Calcomp got an opportunity to tap a large financial source: cash-rich Sanders Associates was willing to buy them up at a good price to extend their presence in the non-military marketplace. Calcomp stockholders, including George, agreed to the sale. George was given a vice-presidency and a position on the board of Sanders. But in nine months, in the summer of 1980, he resigned from both positions to start up a new company "that wouldn't compete with Sanders or CalComp." That company was Novell.
Jack and George, each in his way, is having a vision about the minicomputer peripherals and business computer marketplace. When Jack approached George it was simply to critique his business plan, but as George reviewed it he got more and more enthusiastic. Jack was writing about his vision! He was so enthusiastic he told Jack he'd like to be a part of it.
Jack was delighted. One of the key things he needed for this new enterprise was financing, and here was a man well-connected into the southern California venture capital community and one that could bring a "heavy-hitting resume" to the Novell effort.
The price was high -- George would come in as President -- but the promise of access to million-dollar financing made it worthwhile.
----------------------------------------------------------------
Bio: George Canova
George M. Canova, born in 1931?, was chairman of the board and president of Calcomp. He was principal founder of Century Data Systems Inc. in 1968 and served as it's president until 1973 when the company was purchased by and incorporated into Calcomp a maker of minicomputer peripherals. Prior to founding CDS he was a director of product development for Scientific Data Systems (now Xerox Data Systems) and was earlier employed in engineering and managment positions at RCA and Burroughs Corporation.
George is important because he was Jack's partner in starting Novell and he brought many important innovations to Jack's basic plan with him. Innovations such as adding a personal computer to the product line.
------------
Timeline:
George Canova founds the company and stays with it until March 1982. George is one of the few people who walk away from Novell, never to return or think about it again.
-------------
---------------------------------------------------------------
Bio: Jack Davis
Jack Davis graduated from BYU in 1961 majoring in Business and French. His first job in electronics was working for NCR as programmer. He gained further experince at Hughes, GE, CalComp, Beehive International and TSI before coming to Novell. Jack is important because he was the man who pulled Novell together to found it and later rescued Novell by bringing Novell to Ray Noorda's attention.
The actions of Jack Davis and George Canova in building Novell Data Systems provide a rich contrast to what happened at Novell Inc. under Ray Noorda.
------------
Timeline:
Jack Davis stays with the company until November 1981. He then plays behind-the-scenes roles until Ray takes over in 1983.
------------
Jack and George were thinking on parallel tracks, but tracks were only parallel, not the same, and both were strong-willed people.
The day Jack and George went to incorporate, Jack got his first surprise. He had been using the name Macro Systems to describe this business. While they were waiting for the lawyer, George said, "You know, I talked to my wife about this new company and she thought of a nice name for it."
"What's that?"
"Novell."
Jack thought. "Sounds nice. What does it mean?"
"It means 'new' in French." George said.
Jack blanched a bit. In addition to his degree he'd spent two years in France on a church mission and was quite fluent in the language. "I don't think so, George. There's either nouveau the masculine or nouvelle the feminine..."
"Well it sounds good anyway."
Jack -- having just read in a magazine article saying that the important thing about a business name is that it isn't as important as what a business does -- conceded. "Yes, it sounds good, George."
For a long time after that the company name meant "new" in French when you asked George, and "nothing at all" when you asked Jack.
Novell started in a vacant truck dealership in an industrial park on the north side of Orem, Utah. From it one had a view of three spectacular sights: Mount Timpanogus, Utah Lake and the Geneva Steel works -- largest steel mill between the Mississippi and the Pacific coast.
Jack, Sherrill Harmon and Joe Maroni were the first to set foot in the building as employees of the new company. They started the cleanup and transformation of the building into an electronics manufacturing facility.
Jack started Novell on the vision of manufacturing a multi-CPU minicomputer system, and the terminals and printers for that system, and reselling software for it.
His vision was to build another minicomputer company that would offer more power for less price -- that part of the vision was conventional. He was going to offer accessories for that minicomputer -- that too was conventional wisdom for the time. The two unconventional parts of his vision were that he was going to build those accessories as well as the minicomputer rather than OEM them, and that the minicomputer power was going to be expanded by adding more CPUs to the box.
The minicomputer was to be the heart of this Novell line. It would be based on the Motorola 68000 chip and would be expandable by adding more 68000 processing boards as the customer needed them. This would allow the customer to use the same computer over a very long span of capacity requirements.
This ability to be useful over a long span of capacity requirements was a challenge, in 1980, for all makers of mainframes and minis. A user who started on, say an IBM System 34, could only grow so far before they would have to upgrade to, say, a System 38 or a 4331. When they did they would have to toss out all their old software and much of their old peripheral hardware -- terminals and printers -- as well.
Moving company data from the old system to the new was an elaborate and laborious process too. And, even more threatening in the eyes of the computer manufacturer, once the customer realized he or she was going to have to go through all this nonsense whether they bought the same brand or a competing brand, they were likely to shop around. The manufacturer "lost control" of the client.
Transitioning from one kind of hardware to another was hard to do and fraught with uncertainty for all involved, so a company that could offer products with extended capacity spans would have a competitive edge.
Jack knew the terminal marketplace, so it would be easy to make money by offering a terminal. In his dealings with Dobbs and Woodbury he had developed a rapport with and an obligation to Rusty Woodbury. He called upon Rusty to join the new company to design another Sperry-compatible terminal, and then to design a terminal for the minicomputer.
Every office computer has a printer. Usually those printers are made by a company specializing in printer manufacture and then OEMed to the computer maker. But why? Why not make the printer as well as the computer? Jack decided to capture this part of the revenue stream too. He licensed a dot-matrix printer design from Tritel -- more contacts from his earlier days -- and included it in the product mix.
In Jack's business plan there was no standalone microcomputer or LAN. These products evolved into the business plan after he and George joined forces.
The "terminal computer" (the company's name for a personal computer) became part of the plan shortly after incorporation. George proposed it for two reasons: first, it would be a quick revenue generator that would finance the minicomputer's development. Second, they had a ready customer. A standalone microcomputer would fit more closely than a terminal or minicomputer to what Safeguard wanted as a computer alternative to the One Write system. (more on this next)
"And," in classic terminal manufacturer logic of the seventies, "if we're making a terminal, why not make it a standalone computer?" (Many terminal makers "tested the waters" of the personal computer market in the seventies, but none survived the shakeouts of the eighties.)
The LAN did not appear in the product mix until the summer of 1981. That story is covered as part of how the Superset group came into being.
As Joe and Sherill were cleaning up the building, Jack and George were flying for Philadelphia. No one on the West Coast had been interested in financing this new startup, but on the East Coast Jack had rekindled yet another contact: Adolf "Dolf" Pere and Pete Musser, the owners of Safeguard Business systems and Safeguard Scientific.
-------------
Timeline:
Dolf Pere and Pete Musser are investors through their holdings in Safeguard Scientific and board members from the beginning of Novell Data Systems through the transition to Novell Inc. and into 1991.
-------------
These men had built Safeguard Business Systems and their "One Write" accounting system into a thriving cash cow. They had used an independant franchise marketing approach to develop a sales force numbering in the thousands.
In years previous Dolf and Pete used the profits to diversify into industrial areas such as metal finishing and transmissions. But in 1980 they could see a much different opportunity appearing. They could see that microcomputers used for accounting might soon supplant some of their One Write manual accounting system demand and they wanted to be prepared. This made them interested in diversifying into high-tech -- specifically into microcomputers.
When Jack and George came to them about investing in a computer company, they were very interested. They wanted a system that they could offer as a high-end migration product for their One Write salesforce.
Dolf and Pete knew from earlier experiences that some of their plans and visions weren't in line with what salespeople and customers wanted from Safeguard Business Systems and the One Write product. They solved this friction by forming Safeguard Scientific as the organization for their diversification. Safeguard Scientific was to be the venture capital organization for Novell. Safeguard would invest in and build Novell into a stable and profitable company, then it would spin off its ownership when the shares could be sold for a good return.
Safeguard Scientific was their vehicle for investing in Novell, but the products Dolf and Pete were interested in seeing come from Novell were for the future development of Safeguard Business Systems.
George and Jack talked to Dolf and Pete. The meeting ended with Dolf and Pete committing to financing Novell and with George and Jack getting a check for $250,000 so they could go to Comdex that year.
Going to Comdex in 1980 said a lot about the Novell that was to come.
Novell didn't have a product. It didn't have a plant. It hardly had an organization or money. But Jack felt that establishing a relation with future resellers of Novell product was important, and he felt he was going to find those resellers at the Comdex trade show that November.
The choice of Comdex was significant. This was Comdex's second year and it was still overshadowed by the giant National Computer Conference (NCC) tradeshow. But the NCC tradeshow centered around mainframes and minis and was felt by many in the industry to be a show for "students and tire-kickers."
The Comdex show was directed towards independent reseller organizations, the new wave of computer stores, and centered around microcomputers. Jack felt Comdex might be small compared to NCC, but it was strategic.
So, George, Jack, and Larry Edwards went to Comdex. They put up an inexpensive but visually impressive booth consisting of nothing but hanging Novell banners and lots of chairs. Dead in the center was an apartment-sized refrigerator wrapped up as a Christmas present. That represented products to come.
-----------------
Timeline:
Larry Edwards is a founder of Novell Data Systems Inc. He comes on in 1980 as vice president of sales and leaves in the late summer of 82, midway through the Time of Six Presidents.
-----------------
Jack's hunch paid off: Comdex was strategic, sitting space was at a premium, and the chairs at the Novell booth attracted hoards of foot-weary resellers. The media stopped in because they were curious and they wanted to see more good marketing in an industry segment still dominated by inventors wearing t-shirts and jeans while showing off their latest technology. The Novell reputation started to grow even before the product did.
Having made their debut the Novell team headed back to Utah to plan a blockbuster year for 1981
In 1981 Jack and George's ideas began their transformation into flesh and blood, and metal and plastic. For 1981 the goal was to put flesh in the organization and assemble a team that would start producing revenue.
First of all, Safeguard wanted Novell to lay the foundation for a growing and profitable high-tech company. Dolf and Pete were counting on Jack and George's long and successful experiences in the minicomputer industry to guide them to making that kind of company. Safeguard was looking first at the bottom line. Jack and George needed to make Novell profitable within a reasonable period of time. This constraint affected the order in which products would be developed and brought to market: quick things would come first.
Second, Safeguard wanted a product that would give their Safeguard Business Systems One Write sales force a way to migrate into selling a high-tech alternative or enhancement to their traditional One Write product -- they wanted a way to sell computerized accounting. Their second focus was on integrating a fruit of the computer age into their existing interests.
This constraint would affect Novell's product line by encouraging the addition of a personal computer to the line.
1981 was a big year for Jack. He needed to finish building his founding team. As he finished that he needed to build a company around that team. That company needed to define, design, manufacture and sell profitable products within the year.
And this was no simple single-product company being assembled. It was a comprehensive computer system maker being assembled in a locale that had never seen a comprehensive system manufacture before. Jack's vision was no mean thing, and it was going to take all the Old West Heroism he could muster from his Utah heritage to make it happen.
George wanted to prove he was right. He had told the people at Calcomp that minicomputer peripherals were the market to be in, and they had rejected him. Now Jack had come to him with a business plan that outlined a great way to get into that market. George's goal was another winning company, and whether it was minicomputers or personal computers didn't matter much. What was important was the peripherals business.
George could see that Jack had done his homework in terms of lining up people and making a plan. This would let him concentrate on a role of arranging financing and fine-tuning the manufacturing process -- his great loves. 1981 would be a great year.
From January 1981 to January 1982 Novell grew from ten founders to 120 people. There was a steady stream of new faces coming in to help in all aspects of company operations. Many good people came in with Novell Data System's flood tide of fortune, and left with it's ebb. Many who left then returned on the flood tide of Novell Inc.
Here is a vinette of the people that came to be part of Novell in 1981:
First will come the founders of Novell Data Systems. These are the people who built the stage and then left. The free market system declared that they didn't quite have their act together; they would have to pursue their dreams in another place, and a little older and wiser (which they all did, more on that in the epilog).
After the founders come those who came and left, then came back again to play significant roles in Novell Inc. These are the people who know more than any others the difference between Novell-before-Ray and Novell-after-Ray.
Finally will come those who came and stayed -- the founders of Novell Inc. -- the people responsible for making the "Ray difference" happen.
Through the years, Jack had worked with a lot of computer industry people. He picked from this vast experience base to build the Novell team. For years he had been discussing business ideas with people on an informal basis. Those who believed in his dream and were ready for a change joined him. Some of the people joined late in 1980, the rest came on board early in 1981.
------------
Timeline:
Joe Maroni is one of the Novell Data Systems founders. He is with the company until March 1983 when he leaves as part of Ray Noorda's first reorganization.
------------
Jack drew other manpower besides George Canova from his Calcomp experiences. He knew Joe Maroni from his work at Calcomp. Joe was working at Storage Technology in Denver in 1980, and Jack invited him to head up Manufacturing.
-----------------
Timeline:
Dennis Fairclough is one of the Novell Data Systems founders. He leaves in March 1982 during the "March Massacre".
-----------------
Motorola wasn't the only one excited about the arrival of the 68000 CPU chip. Dennis Fairclough was finishing his PhD thesis which compared the relative merits of the 16-bit processors being introduced in 1980. He showed Jack his work -- which showed that the 68000 was a superior chip to the Intel 8086 or Zilog 8000 offerings -- and Jack invited him to help start Novell. Dennis would be the one who designed the multi-CPU minicomputer.
Dennis was in academia when Jack invited him to be part of Novell, but he was no stranger to commerce. He moved often back and forth between the two worlds. He had previously been involved in the startup of Praxis, a company that sold minicomputer systems with customized software to dentists.
Dennis agreed to be part of starting Novell, but he didn't leave BYU. He still had his PhD to finish and staying at BYU kept Novell's startup expenses down. By the time he was ready to join full-time, Novell was in trouble, so he was involved as a founder, but never as a full-time employee.
The "people conduit" from Praxis included Larry Edwards. Larry was born in southern Utah. Every new acquaintance ran into one vestage of that: his name is pronounced "lorry" rather than the more conventional "larry" associated with it's spelling. His work experience included stints at IBM and Praxis, the dental software company that Jack would head following his term at Novell.
---------------
Timeline
Sherrill Harmon is one of the Novell Data Systems founders. He lasts until March 1982 when he leaves as part of the "March Massacre".
----------------
Sherill was born in Price, Utah. He was one of those people that people are talking about when they complain that "the brightest and the best won't stay" in a particular area -- in this case Price: a coal-mining town in the center of Utah. He left to become part of the electronics industry that was springing up in many places in the West, but not Price, Utah. His journeys took him to Phoenix in the sixties where he joined the GE disk drive division where he became acquainted with... who else... Jack Davis.
In the late sixties General Electric's disk drive division in Phoenix went through one of those economic convulsions that periodically shake companies in the electronics industry. The convulsion leads to contraction and, as is also common in the electronics industry, as GE was contracting, Motorola was growing, so a "people conduit" was established from GE to Motorola. Sherill was part of that conduit.
At Motorola he became a project leader on the 68000 chip development. The minicomputer Jack had in the business plan was based on the 68000, so he invited Sherill to join Novell.
Sherill came, and while the multi-CPU minicomputer envisioned by Jack was never built, it successor, the LAN file server, was 68000 based.
--------------
Timeline:
Rusty Woodbury is one of Novell Data Systems founders. He lasts until March 1983 when he leaves as part of Ray Noorda's first reorganization.
--------------
TSI provided Novell not just with Jack Davis. It also gave Novell a legacy of commitments. One was between Jack and Rusty Woodbury. The handling of the rights selling to the Dobbs and Woodbury terminal had caused acrimony between Dobbs and Woodbury. That, and the company's rocky fortunes, proved sufficient grounds for a split. Jack invited Rusty to join Novell and design terminals -- including another for the Sperry marketplace.
------------
Timeline:
Phil Long is a founder of Novell. He joins in 1981 and leaves as part of the March Massacre in 1982.
------------
Jack called upon Phil because of his experience designing terminals for the IBM mainframe marketplace. Phil came on to head up software design and he became deeply involved in the developing the various low-level programs needed to get the terminal computer going.
------------
Timeline:
Ried Clark is the longest lasting of the Novell Data System founders. He is the only founder to survive the transition to Novell Inc. He retires from Novell in 1987.
------------
There was another connection Jack had made in his wide-ranging activities but had forgotten by 1980. There was a trainer at GE that remembered working with Jack, but not vice versa. In 1980 that person walked into Novell's new office as a representative of another Utah startup computer company -- Billings Computer -- to sell floppy disk drives to Novell. He felt that since Billings Computer had just bought the disk drive operation from Calcomp (part of George's divestitures to raise cash) and George Canova had started that disk drive operation (it had been part of the Memories Division sold to Xerox, but Xerox had declined to take that part so it got sold to Billings Computer in a separate deal), there ought to be a good fit.
When he got there and discovered that his old associate, Jack Davis, was also involved, he asked to join and signed on to head up international sales. This person was Reid Clark.
One hundred people were cast out of Novell during its first time of troubles. Many couldn't stay away. As Ray Noorda reassembled the dream, they came back for a second try. Here are a few.
-------------
Timeline:
Jim signed on with Novell in 1981 to head up technical service. He left early in 1982 as the magnitude of the crisis was becoming evident.
He came back late in 1983 to continue working on service. Jim advanced steadily within Novell and by 1991 is a Vice President heading up Sales of NetWare.
-------------
Jim signed on with Novell Data Systems to head up technical service. Throughout 1981 he labored to set up maintenance agreements with the big service outfits such as Icot. He left early in 1982 to work with Peripheral Systems, another Jack Davis offshoot, as the magnitude of the crisis was becoming evident.
He came back late in 1983 to continue working on service and thrived within the Novell organization until the mid-nineties. Jim isn't one of the "founders" that survived the troubles within the Novell organization, but he's an "almost founder".
-------------
Timeline:
Roger White is about the 20th person to join Novell in 1981 and is laid off when the company shrinks to about 20 people in the summer of 1982. He returns to Novell in the summer of 1986 as Novell tops 500 employees and leaves again in the summer of 1989.
-------------
Roger White came to Novell after spending three years running a Computerland retail store. He brought marketing expertise on reaching the retail store channel, and technical expertise on dealing with the rapidly developing CP/M applications world. Roger started as head of customer support for the personal computer, then moved to sales and headed the Dallas regional sales office for six months, when the "ebb" caught him.
He spent three years at Beehive International working with their personal computer, and then returned to Novell to work in Communications, working on strategic reports and developer relations. He left Novell with the 1989 organizational "phase change."
----------
Timeline:
Dave joined Novell Data Systems in 1981 as an engineer. He left in the summer of 1982 to join Beehive International, but returned in the summer of 1983 as head of engineering under Ray. Dave has prospered well into the nineties in Engineering.
---------
The Novell Inc. founders are the people who survived Novell Data Systems time of troubles. The hard sieve of layoffs and the soft sieve of having the opportunity to pursue other more attractive opportunities had skimmed off 100 people and left these people behind to carry on the dream. Some of the founders continue on in quite prosaic fashion -- they do their job but didn't rise to the challenge of shaping a new company and a new industry. A few, the ones this story is about, did rise to the occaision. Here are the founders we will follow through this tale.
--------------
Timeline:
Craig Burton joins Novell early in 1982 at the beginning of the Time of Six Presidents. He stays with Novell until March of 1989 when his departure, along with that of Judith Clarke's, signals the beginning of the Organizational Phase Change of 1989/90.
--------------
In 1977, Craig Burton stepped off the plane and took deep breath of the dry, thick air of home: Utah. He was still over 4000 feet, but he'd come from a mission in Ecuador where he was riding a bicycle and proselyting for the Mormon church at over ten thousand feet. He listened to the airport speaker. It was English, not the South American Spanish or the native Indian Quechua he'd been hearing for the two years previous.
It was good to be home, but what to do now?
Craig started working with his father, LeR Burton, selling real estate. But the income was slow in coming. In fact, he only sold one house. But real estate was only one of many things that interested these two Burtons. Computers and electronics was another. LeR, for instance, had worked in the middle seventies with some of the first Hewlett Packard handheld calculators, the HP-35's, and had written some of the examples that HP put in their instruction manuals for doing real estate calculations.
Craig and LeR started doing computer consulting -- helping businesses integrate personal computers into their work. They started working with Roger White at Computerland as a supplier, and they got a couple Apple II's. Craig was soon programming cash flow projections using Apple's Pascal programming language.
By 1981 Craig was out of real estate entirely and into personal computers entirely. He was working with Personal Business Computers (PBC), a thriving retail computer store in Salt Lake, to sell systems to professionals, and he was discovering that while professionals liked the pricing and the office productivity software available on personal computers, they needed more connectivity than personal computers offered.
When Craig saw the first ads for the Corvus computer networking system, he thought he was seeing a godsend. Then he tried to make one work. It didn't. He called Corvus many times and even visited their factory once. He wasn't satisfied. By 1982 he knew they were on to something hot, but he could also tell that they hadn't the foggiest idea what it was they were making, and it would be a while before they found out.
By 1982 Craig knew a lot about selling personal computers. He had participated as Apple and the various CP/M-based computers had battled for the hearts and minds of computer retailers and their customers. He had witnessed Commodore shoot themselves in the foot time and time again with ham-handed marketing and incompatible product introductions, then save themselves time and time again with price reductions that opened markets of new buyers and new store operators who saw those prices and said, "Lets start here."
He had witnessed Apple bobble the Apple Three introduction -- in part by ignoring how important the Apple II's "open bus" had been to its attractiveness. And in 1982 he was watching them compound the error with their Lisa product introduction by making it both closed and pricey. (Lisa was the short-lived Macintosh precessor, that taught Apple how not to market Macintosh.)
Craig had also been noticed by several people at Novell. He was interested in local area networks at the same time Novell was interested, and in spring of 82 he got his break. Roger White was moving out of customer support into sales and he needed a replacement. Craig came to mind. Roger talked with him; he visited; talked with top management; expressed his views on LANs and was immediately invited to speak at a seminar in St. Louis.
Weeks later he was formally on board doing customer and sales support work.
Craig did a lot of things at Novell but he spent most of 1982 working with Reid Clark on building international sales. He and Reid traveled worldwide. As a result they were partly isolated from the crises that shook Novell in the "time of six presidents."
As the crises ended, Craig was working on documentation, customer support and, in what ended up being his most strategic effort, developer relations.
--------------
Timeline:
Harry Armstrong joins Novell in the summer of 1981 to work for Joe Maroni in Manufacturing. Harry survives the Time of Six Presidents and heads up both Manufacturing and Sales under Ray Noorda. He stays with Novell until summer of 1987, when he retires to take up ranching on the Idaho/Wyoming border.
--------------
Utah has seen it's share of electronics failures as well as successes. There is a straight-as-an-arrow section of I-15 at the south end of Salt Lake valley just before the highway winds and curls to snake by "point-of-the-mountain" -- the landmark hill that separates Salt Lake valley (which holds Salt Lake City) from Utah valley (which holds Orem and Provo). Midway along this stretch of highway is a giant brick building that houses the Utah National Guard -- easily identified because out front is a jet fighter mounted on a forty foot high pedestal.
This spectacular building is the legacy of a multi-hundred million dollar electronics industry failure. The builder of that building had dreams of churning thousands upon thousands of computer terminals out of it. The builder was ADDS corporation, a New York company that in the mid-seventies was one of the leading CRT terminal makers, and Harry Armstrong was one of the people that was going to make that happen as an employee of ADDS corporation. But ADDS suffered the same declining-price-and-declining-margin challenges that faced Beehive, Novell and all the other domestic makers of CRT terminals in the late seventies and early eighties, and even Utah's low labor costs weren't enough to keep them competative with the East Asian imports. Shortly after it was built, the factory was shut down.
At ADDS Harry was a "metal bender" and "plastic pourer". He was in charge of making the physical manufacturing processes happen, and it was this expertise that attracted him to Joe Maroni. But Harry was another of those people with a rich background in electronics. Prior to working for ADDS, he had been one of the founders of a personal computer manufacturing company in Ogden, Utah that struggled, but didn't make it, and prior to that he'd been in California doing design and troubleshooting work on flying survellance systems used by the military in Vietnam.
Harry enjoyed working with people, and Novell turned into a perfect stage to exersize his many people and technical talents.
--------------
Timeline:
Judith Clarke starts working for Novell part-time in the summer of 1981. She becomes a part of Novell full-time as Ray Noorda comes on in 1983. Under Ray she heads up Communicatons. She stays with Novell until March of 1989 when her departure, along with that of Craig Burton's, signals the beginning of the Organizational Phase Change of 1989/90.
--------------
Judith's story is an "American Gothic" of the eighties and nineties. She grew up in humble backgrounds in the period between "Rosie the Riviter" of WWII and the "Woman's Lib" of the seventies and eighties. She was outgoing and ambitious, and she was constantly driven to make something of herself in a time when this ambition wasn't taken seriously in the average woman.
By 1981 she had married twice and born three children, but there was little one could point to on her resume to say, "giant success awaits you just around the corner." In 1980 she was in Hawaii working part-time with her husand, Reid Clark, at a company that sold minicomputer-based energy management systems, and before that she had done some presentation work for a small telecommunications company.
Judith met Reid at a church social in Phoenix, AZ, and they married shortly thereafter. She followed him as his work took him from Phoenix to California to Hawaii and finally to Provo, Utah.
-------------
Timeline:
Superset is a group of sometimes three and sometimes four programmers who did the pioneering development work on the NetWare LAN operating system program and the various improvments to it. They came to Novell as contract programmers in the summer of 1981; formed into a distinct entity in the summer of 1982, and continued working for Novell as such into the nineties.
-------------
In 1981 Jack Davis was looking for a word processor to offer as "Novell's word processor", and in 1981 BYU was a hotbed of word processing activity. It had a core of graduate students and professors with a serious interest in word processing, and this core had been developing word processors for the last two years for organizations such as Hill Air Force Base, Savin, and Qume.
Jack consulted with an assistant to the Dean at BYU's College of Engineering. He recommended four programmers to Jack. (One of whom, Dale Neibar, had done work on a word processing project for Jack the year previous -- yet another of Jack's enterpreneurial projects.)
("Blistering Hotbed" is the proper description for BYU that year. While four programmers of the word processing development community headed off to make their destiny happen as Superset at Novell, another group started what would become WordPerfect Corp.)
The programmers were mustered, but before they could work on a word processor, they were directed to a more pressing need: a product introduction for Comdex.
The story of Superset -- which will come later -- reveals a lot about the functioning of Novell management before and after Ray Noorda came on board.
Jack's list of connections didn't end with people that joined Novell. Many helped Novell without actually joining it.
One example is Sherill Harmon's superiors at Motorola. They were also associates of Jack's. Even though he was "taking" one of their people back to Utah, they saw his work as a potentially big market for their chips, so they remained enthusiastic about Jack's work. They helped Jack and Sherill stay up to date on the 68000's progress. Novell was usually the first or second company in the state to get update information and newly revised 68000 chips.
Another person who was aware of Jack's work, but not involved, was Ray Noorda. Jack and Ray worked at GE in Phoenix. In the seventies Ray moved on to become president of General Automation and Jack worked for him there, heading up international sales.
Jack presented his plan to Ray, but Ray was involved in turning around Bochart Industries, a maker of power supplies for electronics. He declined to get involved.
Ray had another reason for staying out. Since he had become head of a company producing a component for computers, he gained a new perspective on industry trends. When Jack talked about manufacturing hardware in Utah -- minicomputers, terminals and printers -- Ray felt he was in for some tough sledding. Yes, Utah was a low wage state in the US, but in 1980 terminals and printers were just the components East Asians were getting good at making. This meant increased competition and declining margins for all involved. In fact, one of his major goals for restoring profitability to Bochart was to move power supply manufacture off-shore. Ray decided to pass on the opportunity Jack offered.
The goal of all the people joining Novell was to build computer products that the marketplace would buy. Jack had an impressive list in his business plan and 1981 was the year in which those ideas would be transformed into metal and plastic.
The first product out the door was the printer. Incorporating the Tritel design had gone quickly, and Image 800 printers were moving off the shipping dock in March. The Image 800 was a medium-duty, wide carriage printer. It printed at 160 characters per second and had a character selection comparable to the early Epson printers: four printing widths, double high, double wide, international and extended character sets.
Through most of 1981 the printers sold well, and this brought assurance that Novell was on the right track. But printers carry low profit margins, so Novell was going to have to sell more than printers if was to sustain the kind of company Jack and George had in mind.
Rusty Woodbury came on early in 1981 and started work on designing the terminal. Jack's vision was that this would be a terminal to sell to the Sperry-Univac marketplace, but shortly after Rusty arrived he was diverted from designing a Sperry-compatible terminal to a terminal that would be the front end for the personal computer. In theory this would be no problem. In practice, Rusty took all the Sperry-compatible ideas floating around in his head and put them into the personal computer product. The result was an engineer's delight: a product with a lot of features that were easy for an engineer to implement, but not particularly relevant to the market for which the product was designed.
The terminal Rusty produced had graphics capabilities and two kinds of programmable function keys. It was housed in a large case and the keyboard was broad and detached from the screen. All this was being presented to a personal computer marketplace that was used to working with simple, small, character-only terminals.
The terminal was premium, but how much of the market was willing to pay for a premium terminal? No one knew and top managment didn't have the time to find out. They were relying on Jack's experience in the marketplace to lead them through to profitability. Once they were profitable, then they would have time to find out. Meanwhile, Jack was busy building a company.
As Rusty finished the terminal for the computer, he turned back once again to working on the Sperry terminal, but before he could finish, the first wave of crisis in the fall of 1981 put it on the shelf.
The Terminal Computer was a child of the Golden Rule: he who has the gold, rules. The Safeguard connection drove this product to the front of the development list. It was a combination of Rusty's terminal plus a 64KB RAM, Z-80 based, computer board. The computer would run CP/M operating system and have two serial ports for talking with printers, modems and other peripheral devices.
The terminal computer suffered from being brought into the planning in a hasty way. Novell was structured so that it was counting on Jack's experience in marketing to substitute for batteries of consultants and hours of market research -- the company didn't have the time for those if it was going to be quickly profitable.
The flaw was that Jack hadn't paid that much attention to the personal computer marketplace. His attention was focused on terminals, printers and minicomputers. In these areas his instinct as to what constitued a winning product was sharp. In personal computers it was fuzzy, but the terminal computer was to be the second product out the door. As a result, he let Rusty's engineer's delight slip out the door without bringing it sharply into focus with what the market was looking for.
The Novell terminal computer ended up being a high-end, feature-loaded personal computer system, when all the company needed to produce was a computerized accounting system that One Write salespeople could sell. But by the summer of 1981, that was what they had, and that was what they had to sell.
Ironically this engineer's delight was a vital step in developing the local area network product.
The minicomputer was to be developed with profits from the sales of other Novell products. It went on the shelf almost as soon as Safeguard financing was arranged.
Jack's work at TSI also garnered him rights to a program called AutoGen. AutoGen never became a Novell product, but its evolution gives more insight into Jack's style.
The AutoGen idea started as a program generator program put together by a group in Hawaii for the Nova line of minicomputers. The intent of program generators is to allow relatively inexperienced programmers to produce code quickly by automating parts of the code generating process. A programmer, for instance, builds a screen with a screen editor, then the program generator scans what the programmer has done and generates code that will tell a high-level language -- typically COBOL, BASIC, FORTRAN, or Pascal -- to build a similar screen. This leaves the programmer free to concentrate on other parts of the programming task.
The program generator concept has been around a long time, but it is one of those parts of computerdom that never quite lives up to expectation. The concept is distinctively cyclical. It rises suddenly as a panacea for each new generation of computer programmer "wannabe's", then dies quickly as those people see the reality behind the concept: progamming is still tedious.
Jack saw this Nova program generator and felt the concept was a breakthrough. He wanted to buy the rights to this package and sell them through TSI. He was foiled. The programmers had already sold interests to other companies and one of those ojected that selling to TSI would conflict with their interest.
So Jack did an end run. He brought in a programming friend of his, to look at the package and asked him if he could reverse engineer it. "Of course." said the programmer and he developed AutoGen for TSI -- Jack put another piece in his "hope chest" for the upcoming minicomputer. (In the tight circles of the Utah computer industry few people disappear. This programmer was Charles Burgoyne -- a computer-related jack-of-all-trades who later started one of Salt Lake's most long-lived retail computer stores).
AutoGEN shows that Jack was resourceful and his connections were widespread through the Utah and southern California computer industry. But he was sloppy about his follow through, he'd build fast, but he'd leave lots of holes behind as he did so.
Now that the stage has been revealed and the actors introduced, lets see what happened in 1981. This is a series of vignettes of things that happened at Novell that Roger White saw. They show the excitement, the innovation, the compromise and the coping, that went on as Novell tried to make dreams fit reality on a limited budget.
Jack Davis brought on Larry Edwards in November 1980 as vice president of sales. Larry's job was to organize those who would sell Novell products. Larry came on early, he was one of the attendees at the first Comdex. This is significant because it shows that from "day one" Jack was bent on selling things. There was no low-overhead "development" phase at Novell, during which product was developed without the pressure of a sales staff waiting expectantly to rip the finished work out of the developers' hands and begin hawking it. Larry knew how to sell minicomputers. He immediately set out to build a national sales force with offices located regionally across the nation.
Larry started bringing on the sales force. He was a man with experience, so he brought in people he was experienced with. One of the first salespeople on board was Andy Olson, who would handle the southern California area.
Andy was born and raised in Northern Minnesota. His dad was a mining engineer who worked at exploring and developing the open pit iron mines in the area. But digging giant holes in the ground wasn't a legacy to be passed from father to son.
Ten to fifteen years ago Andy reached the California coast and flourished. When he signed on with Novell he had two houses and two ex-wives in Manhattan Beach, and he was something of a local celebrety. He was into supporting local politicians, and he was trying to reach a final settlement with his second wife. And he was also a remodeling contractor.
He was the kind of guy who was always busy and always on the phone, except at lunch when he went to Orville's, a classy restaurant in the center of Manhattan Beach were Andy knew the owners, and the banker that financed the owners,.... And on it went. He treated the computer industry the same way he treated Manhattan Beach. He was one of the speakers and advisors to the first Comdex -- Comdex 79 -- and he had connections all through the southern California minicomputer sales scene.
Andy's experiences with introducing Novell products were typical ones.
-----------
Timeline:
Andy Olson is one of Novell's first salespeople, he comes on in March of 1981, and one of the first to see how fundamental its problems will be. He leaves in May of 81 shortly after the first personal computer is shipped.
-----------
Roger was brought on in March to head up customer and sales support, but since there weren't any customers yet, he helped in the task of getting some. Roger coordinated with Andy to set up a series of meetings with the five Computerland stores in Southern California.
The meetings went smoothly enough but the response was reserved: none of the store owners wanted to be a pioneer. They wanted to see product and see customers before they committed.
As they left, Roger hummed "The times they are a changin". The days when a Computerland store owner asked only two questions: "Does it run CP/M?" and "Will it work when I turn it on?" were ending quickly.
Andy moved on to Plan B.
Andy's next stop was Jade Computer, a mail order house pioneering in the offering of personal computers. Andy had connections there. The Jade people were more responsive. Mail order, it seemed, was still in the two question era, and Jade was signed up as Novell's first customer.
This left Roger uneasy. Mail order was the low-end, whorish, competitor to the retail store channel. It wasn't clear that taking on Jade as a reseller would make the task of penetrating the retail channel any easier. It might, if it helped generate awareness and demand, but it might not if it started a trend of margin shaving.
What Andy knew that Roger didn't was that Jade had a retail side as well, and this made it look attractive to Andy.
So, Andy had his first blood and Roger returned to Utah.
The quickest way to turn Rusty's terminal design into a computer was to add a CPU board to the terminal enclosure and put the disk drives in a separate box. This was the path chosen, and it was a good one for several reasons: the terminal enclosure was big and heavy already, so having the disks in a separate box kept the shipping and handling weight down. In the early eighties, disk drives were still noisy. With a separate box, the drives could be moved away from the operator.
There were also pitfalls in this simple approach. The pitfalls had to do with how these various parts were going to communicate with each other.
The terminal part of the microcomputer was already designed to talk with the outside world. It was equipped with the twin RS-232 ports that made that possible. Once again, the simplest solution was to use those ports and let the CPU board do all its communicating through the terminal board. But this was a deviation from standard microcomputer practice in the early eighties. The terminal computer looked like this:
------------
|terminal |------ RS-232 to outside world
------------
|
---------
| CPU |
_________
rather than this:
------------
|terminal |
------------
|
---------
| CPU |------ RS-232 to outside world
_________
The put-a-CPU-behind-the-terminal approach produced several other differences such as this that presented subtle challenges. Solving these challenges added overhead to the system, and in spite of its size, and the impressive claim of four CPU's in the system, its performing speed was mediocre.
Rusty was brought on board to design a high-quality terminal that supported graphics. When the terminal turned into a microcomputer, there seemed no reason not make it a graphics-supporting microcomputer, so he did. The decision to go with graphics produced another performance hit. The screen generator didn't use a fast character generator chip, it relied on slower ROM-based software.
One place that the terminal computer did excell was in offering a hard disk. In the early eighties, attaching a hard disk to a microcomputer was not unheard of, but it wasn't common, either. Novell took the bold step of making hard disk an integral part of the system: you always got a hard disk with your Novell computer.
One of the first things Jack and Larry needed to decide was how to approach their marketplace: the newly emerging retail computer store channel. Jack knew such a channel existed because these new stores were buying some of the terminals that he sold at TSI. Larry in his Praxis experience had few dealings with them since Praxis software was sold on minicomputers.
One of their early sources was an ex-Computerland store owner who joined the company a couple months later: Roger White. With his input about how Computerland was organized and what was expected of margins and pricing the product marketing and promotion, features were hashed out.
Roger recounted his pricing experiences to Jack and Larry, "In the early years of retail computer stores -- the late 70's -- hardware products carried a margins of 25 to 33%. These were very low. One of the first lessons in the Computerland training seminars was a cash flow lesson that demonstrated how important a good margin was to profitable sales. The moral: don't discount just to increase sales volume. If you do discount, make sure you pick your margin back up on the accessory sales.
"One of Computerland corporate's major goals was to raise that margin, and by the beginning of the eighties the 25% items were rarely seen and a few 40% items were appearing."
Given this background the terminal computer was priced with a 45% margin and no discount for volume dealing. Novell wanted something dealers could become strong selling.
As Larry assembled the sales force, it was Roger who trained them. He assembled classes in CP/M basics: booting the system, word processing with Wordstar, and showing how these tools could be used in office automation.
Novell was going to build a CP/M computer, but no one in the front office or top managment had any experience with one. The only CP/M machines at Novell were two Intertech Superbrains in Phil Long's lab, and they were in the usual state of Superbrains -- dead. (Intertech was a company that survived for several years on the new blood entering the personal computer field. Their quality was terrible, but they were always the lowest price, so they always got tried first.)
Roger saw this as a big oversight. To give the company some insight into what it was dealing with, he brought in his pride from Computerland days: a dual eight inch drive Digital Microsystems computer, with an IBM 3101 ASCII terminal, and a Diablo daisywheel printer. He also brought in his software collection, and set the system up near the break room, where lots of people could have access to it.
From this system, the company managment and staff learned their first lessons about CP/M word processing and database management.
Novell, following minicomputer tradition, was going to sell software for their personal computer. Roger, being the most experienced with personal computer software, got involved in picking packages.
Novell needed a word processor. The most popular in the CP/M world of 1981 was WordStar. Novell signed a deal with MicroPro and began offering WordStar.
It was a good choice. Wordstar was a well-known, full featured package that worked well on the terminal computer. With a little extra work, it was customized so that it could take advantage of the function keys on the terminal computer as well, and it became a hot package.
Like the course of true love, the course of true business rarely runs smoothly. Jack felt that WordStar wasn't a Novell product. It was okay as an interrum solution, but Novell needed something that Novell could call it's own, and get better margins on. He continued the search.
This resulted in a parade of word processing developers coming in to offer their product as something Novell could OEM. Jack also talked with an assistant to the Dean at BYU about what features the Novell word processer should have. This fellow contracted to do some research for Jack and brought in some BYU computer science grads to help him. This was the start of the event sequence that brought the SuperSet programmers to Novell.
"And while that was happening, other big changes were occuring. Jack and Mike King's search for "novell's own" word processer was bearing fruit. A group from Phoenix called Metasoft had reached Jack's ear via Sherill Harmon. They would build Jack a word processer. As credentials they presented their existing WP(word processor) -- Benchmark. They had sold this package to Zenith and Harris.
"I had seen benchmark in its Zenith incarnation and it was no great shakes. It was page oriented and did nothing special in my mind to compensate for that. Wordstar was better.
"But Jack wanted Novell's own, so he struck a deal: we would give them a machine, they would give us a WP. Like many other deals Novell made during this period, it sounded simple enough, but the implimentation was another matter.
"In this case we had a chicken and egg problem. Metasoft needed the latest terminal firmware to develop their software on so it could incorperate all the bells and whistles they had aggreed with Jack to add. We didn't have the latest firmware released yet. But Jack had a show coming up and he wanted to show the WP off.
"So Jack put pressure on Rusty, who would then release preliminary (and buggy) firmware to Metasoft. But, as soon as Metasoft found bugs, they would stop work.
"At one point I was appointed project manager to see if this could get completed. But there was no way I could get into that loop meaningfully, and after a week or so no one tried anymore. I didn't because I became a sales manager and all I needed was Wordstar. "Novellwriter" was a waste of time as far as I was concerned."
In 1981 the Visicalc spreadsheet program had been out for two years. It was immensely popular on the Apple II, and had made a business machine of that computer. But Visicorp was very slow about porting the package to other kinds of computers. They had a Commodore PET version out, but nothing for the CP/M world. Roger White called them several times and always got back a "we're considering it" sort of answer.
Roger knew having a spreadsheet was as critical as having a word processor. For weeks he phoned, researched and network through his connections in the personal computer world looking for a suitable substitute. Just as the impossible looked like it was happening -- there was no substitute -- the Osbourne 1 computer was announced, and bundled with it was a spreadsheet package called Supercalc from an outfit called Sorcim Software (Sorcim later merged with Computer Associates). The Osbourne 1 was a CP/M machine. Roger called Sorcim and Novell was probably the first place in Utah to have Supercalc -- the spreadsheet problem was solved.
Roger didn't see any way Novell could hope to be a single source supplier of applications software -- the personal computer world wasn't expecting Novell to be one, either. What they were expecting was an easy way to get CP/M applications from many sources onto the Novell disk format.
Some Novell personal computers used 8" drives -- which were fairly standardized, so getting applications on 8" disks that Novell computers could read wasn't a problem. But the 5" drive situation couldn't be solved so easily. The problem was that there wasn't a single standard for 5" drive disk formats, there were dozens.
There were two alternatives: either Novell was going to become a center for disk conversions or it was going to offer a product that could be sold into the field. Then conversion would become a reseller activity -- a much faster, more flexible solution.
The solution Novell ultimately offered was an error-checking file transfer program developed by Gary Byrom, another Utah pioneer in the personal computer industry. This product, BSTAM, could communicate between the serial ports of different kinds of CP/M machines. A reseller could buy 8" versions of software and use BSTAM to transfer the programs to Novell 5" disks.
Word processing, spreadsheet, file transfer and terminal emulation. These were all areas where Novell acquired products that were popular, workable and easy to install. Then came the "great Russian front" of personal computing, and Novell in particular, the accounting packages.
One of the hallmarks of personal computing in the eighties was its simplicity. The standalone personal computer system was simple compared to a minicomputer. It consisted of a CPU, terminal and disk system. The operating system was simple -- one computer and one user. Even buying one was simple -- just walk in a retail computer store and buy one.
Most application software was also simple.
You want to word process? You buy a word processing package.
You want to spreadsheet? You buy a spreadsheet package.
You want to program? You buy a language package.
Then comes accounting...
The difference between accounting and the other commodity applications is you can't tell someone how to account without telling them how to run their business. Worse, computerized accounting tends to be very deterministic -- it's hard on fuzzy, seat-of-the-pants styles that are characteristic of small business people who are attracted to personal computers rather than minicomputers. Worse still, people ask for (or at least they're impressed by) accounting packages designed by CPA's -- who are the most deterministic of the people who deal with accounting. (It would be better if these people asked for accounting designed by a housewife who tends three kids after school!)
The confluence of these trends resulted in a marketing challenge in selling accounting packages similar to that faced by the life insurance and funeral plot industries: selling something that's considered important, but rarely used. Most of the buyers of personal computer accounting packages bought them as justifiers for getting the computer, but used them only briefly. They would use the computer's word processing and spreadsheeting pretty much as they had anticipated they would, but they would use a hybrid spreadsheet-manual system to accomplish their accounting goals because the accounting packages were too procedurally rigid to suit their needs. The accounting package would sit on the shelf beside the BASIC programing language package and the user would mutter, "Some day, I'll get back to that and make it work."
This use dynamic made the marketing and presentation of accounting packages more important than the content. One of the first to discover this was Tom Byer, the driving force behind Peachtree Software in Atlanta, Georga. While other personal computer accounting houses were selling their product in plastic baggies, he started offering handsome vinyl binders, and spent big money on offices, customer support and advertising.
One group he attracted was top management at Novell.
They wanted to offer an accounting package, and, like WordStar, Peachtree had a reputation as something that would fill the need. The decision was made to carry Peachtree.
Roger got the shudders when he heard that Novell would be supporting any accounting package.
The first purchaser of a Novell Terminal Computer was Novell's landlord, K&P Plumbing. They wanted to computerize their accounting and Jack offered them a Novell system. The following is an excerpt from Roger's diary.
"With the shipping of the first units, my job [of customer support] came on line. The first unit we shipped was to our landlord, K&P Plumbing. He bought it direct from Jack Davis at a discount. Jack sold him not just the hardware, but a complete accounting system -- and we were going to install it. Trouble City!
"Gary Dan Hill and I got saddled with the installation. We were to install the full Peachtree system -- GL, AR, AP, Payroll and Cost Accounting.
-----------
Timeline:
Gary Dan Hill is a Novell two-timer. He comes in 1981, leaves in 1982. Comes again in 1984 and leaves in 1987.
-----------
"Dreams, Dreams, Dreams! Peachtree was just converting from Series 4 to Series 5. All the packages were buggy. Payroll wasn't ready yet and, it turned out, it wouldn't be for months. There was no cost accounting planned for Peachtree [which K&P specifically needed] and no easy way to do it -- Jack had not checked before he said, "Yes." Jack was also reluctant to sell K&P anything but Peachtree since this was to be "our" accounting package. The hardware was also brand new so it was somewhat troublesome.
"Finally, the K&P people had hired a new girl to run the system and this was her only job. She was an experienced operator on mini equipment, but new to micro's and CP/M. Worst of all, she was new to K&P, so her job depended on the performance of the computer.
"The net result was six weeks of constant attention from Gary and I. We were constantly back and forth between Sandra at K&P, and Janice and Janet at Peachtree in Atlanta, and Jim Bills at HQ, trying to get things diagnosed and fixed.
"We learned Peachtree, Oh boy, did we learn it! I was also totally discouraged at the stupidity of our getting involved in supporting an accounting installation at an end user site.
"It was K&P that drove me out of customer support. I told Larry I needed to make more money and asked him what was the best way. 'Go sales, young man, go sales', he told me (I paraphrase). So I began negotiations on getting a position as sales manager.
"In the meantime, the heat at K&P finally ended when Sandra gave up. She quit and K&P was slow to refill her position -- the system gathered dust."
The problem Novell faced in selling Peachtree accounting was, ironically, that it was selling a big, expensive personal computer. The personal computer had a hard disk; it had function keys; it had a classy case. Purchasers expected more from it. It wasn't the forgiving "toy" that an Apple II was. So, while Peachtree could adequately serve the TRS-80 and Apple II marketplace where few users would actually use the product, it wasn't going to meet customer expectation in the Novell marketplce.
In a crisis a person's heart rate quickens, the breathing deepens and habit short-circuits judgement as a way of making decisions. Peachtree was a crisis at Novell, and with the shipping of the first of the terminal computers, top management was finally getting some real, and harsh, feedback about their creation.
Top management's "habit" turned out to be calling on outside consultants to get a "reality check." Habit also said, "Do this by consulting with someone you trust." Since George was the top manager, the consultants brought in were people George, not Jack, trusted. Unfortunately, the people that George trusted were even more deeply steeped in minicomputer practice than Jack's associates were.
The habit that replaced judgement was to turn to a minicomputer-like solution from the "good old days." This translated into "BASIC is a bad platorm for accounting; COBOL is a good platform." Peachtree was uncerimoniously dumped. The excuse being: because it was running on an interpreted BASIC platform and because it took too long to switch from one module to the next (about 30 seconds).
In the meantime, George's consultant found a COBOL-based accounting package that had been ported to CP/M. He recommended it to George, and George declared it would be the new Novell accounting package.
This new solution had two problems: first, bad programs can be written in COBOL just as easily as in BASIC. And, in the preliminary assessments of the package made by those who bothered to run the package, this was amply demonstrated.
Top management didn't listen easily. Novell was just about to sign on the dotted line when Robert Lundahl -- newly hired into the marketing section -- suggested to Larry Edwards that he watch this new package do a close. The close took an hour. Larry got the message; he showed the other top managers, and it was finally back to the drawing boards once again.
The second problem was this COBOL accounting fiasco opened up serious questions about managment competance. The old saying, "Fool me once, shame on you. Fool me twice, shame on me." was in the mind of Novell resellers. Novell management had goofed twice trying to pick a package.
It wasn't just resellers that were upset by this series of accounting fiascos, either. Roger had decided to depart Customer Service because of the Peachtree experience. He moved into Sales. Robert Lundahl decided to depart the company entirely because of the accounting experience.
Robert's outrage came from a deeper source than Roger's. He had written an accounting package of his own in COBOL. When he heard that Peachtree was going to be dumped in favor of a COBOL-based package, he offered his for Novell's consideration. On the advice of George's consultant, they refused it -- faulting it as not comprehensive. When Robert saw the trash that the company had almost accepted, he left -- stating he couldn't quite follow the company's logic on why they would consider this deficient out-of-house product, but not his. Robert sniffed the airs around this situation: it smelled. The accounting crisis hit its nadir.
Habit hadn't worked, but the next step top managment took did: they ignored the problem. It went away. It turned out that in the world of personal computers, it wasn't necessary for the hardware maker to provide application software. In the personal computer marketplace, that was the role of the reseller. The reseller could contact applications developers and distributors as easily has the hardware makers could, and, as long as the operating system was standard, he could install it just as easily too. The simplicity, low capital requirements and standardization of standalone personal computers drove a change in marketing channel expectations. It made efficient system integration possible at the reseller level rather than at the hardware maker level.
Novell stumbled on to this truth, but it was a hard one to recognize because other events were commanding most of the company's attention.
Dolf Pere and Pete Musser invested in Novell for many strategic reasons, such as it should be very profitable, and they wanted to migrate the Safeguard Scientifics company into high tech. They also had a tactical reason: they wanted something to offer their Safeguard Business Systems organization an automated successor to the manual One Write package they were currently selling.
The man charged with making a terminal computer variant suitable for the Safeguard Business Systems marketplace was Phil Long.
Phil Long was in charge of software development at Novell. He was a smart, blunt man who liked talking about the intricacies of the technology. "He was the kind of guy who if you asked for the time of day would preface the discussion by telling you the art of clock building." said Reid Clark.
He was also very conservative in his preference for technological styles, which often put him at odds with Rusty Woodbury who was designing the hardware. One such conflict was over the protocol used to have the terminal board talk to the CP/M board in the terminal computer.
Some background: The terminal computer was literally a terminal board plus a computer board housed in the same casing. The two boards had to be connected with appropriate circuitry and a cable to make communications between the two boards happen. The choice of hardware connection was curious to start with: given the short distance between the two boards and the fact that Novell controlled the circuitry on both ends of the link, a parallel connection would have worked quite well. It would have been a fast, reliable, inexpensive connection. But these boards were a product of their own histories, so the link was made serial instead -- a slower, more expensive linkage, but one for which the circuitry on both boards was already in place.
But this expedient choice would dog the terminal computer throughout its life because it aggravated the slow screen display problem. Further aggravating this problem was Phil Long's choice of how to handle the communications protocol -- and this is where he and Rusty locked horns.
Phil said something like, "This is a serial connection, and an important one, we'll handle communication over this link with a formal handshaking and error correction protocol."
To which Rusty replied something like, "Using a formal protocol is going to double our overhead over an already-slow link. That formal protocol was developed for modem use where lines are presumed to be noisy. Lets presume this line is clean and use an abbreviated protocol that'll double our throughput. We'll add just enough error checking to tell if we've had an error. If we get an error, we'll recover and retransmit."
This was the "Phil and Rusty relation", and dealing properly with this kind of relation is one of the high arts of keeping a company two things: alive and creative. In a healthy company, these differences are called "creativity at work" because they insure that all sides of issues and all relevant issues are examined. In a company that's on the rocks, these differences are called "interdivisional communications problems" and are said to be the root of the company's inability to face the crisis.
Both views are correct. The difference is making sure that these people are communicating and making decisions before, not after, the finger-pointing stage. One of the keys to doing this is establishing a good company vision.
The high point of Novell Data Systems was the Summer National Computer Conference (NCC) 81 show at the Mc Cormic Center in Chicago. Novell wasn't planning to attend until the last moment -- a couple months before the show. Once the decision was made, the company went into high gear planning how to make a memorable impact on short notice and short funds. Novell's showing at this NCC was an example of Jack's promoting at its finest.
The first problem was even getting a booth. The waiting list by spring was 150 companies long. If a booth couldn't be found, Novell would have to base from a hospitality suite at one of the hotels rather than be at the convention center -- which wouldn't be nearly as effective. Thanks to Andy Olson's connections, Novell got a 10 by 20 booth on the main floor when Jade Computer decided to back out of the show. A display was designed in short order.
Then Jack came up with his coup-de-gras. He swung a deal to pay the chartering costs for some of the shuttle buses moving convention attendees between McCormic Center and the hotels. In return for covering their costs, he ran low-key tape-recorded Novell ads through the bus intercom. "It's a captive audience." he chuckled. In addition, when the attendees arrived they were handed a newspaper insert piece describing Novell.
The show went quite well for Novell. The printer and the terminal computer were displayed and the LAN was announced; interest was high.
NCC 81 was the start of Judith (then Judy) Clarke's (then Clark) involvement with Novell. Judy was Reid Clark's wife. For months she and Reid had been conspiring to get her involved in Novell. Earlier Reid had proposed taking Judy on one of his business trips to Europe as his secretary. Jack had vetoed it. As NCC approached Reid proposed bringing Judy along as part of the exhibit team. Once again, Jack was dead-set against it, "It would set a bad precedent -- having a husband and wife work at the company." Jack's wasn't an uncommon feeling among businesspeople, but it made more sense a decade earlier, and in communities larger and less tightly-knit than Utah Valley -- there just aren't that many choices of places to work. And, in spite of Jack's protests, it was already happening at Novell.
But as the show deadline neared, the booth shipping deadline was missed.
"The only way to get it to Chicago was to load it in a rented truck and have someone drive straight through." relates Reid. "I'll do it, I volunteered, but Chicago's 20 hours away. I should have someone go with me. How about Judy?" Jack relented. And so it was that Judy started working for Novell. She drove, and then she worked at setting up and tearing down the booth, and took her share of booth duty as well.
But the road wasn't smooth for her. She was in again and out again at Novell several times after that NCC assignment. She and Reid would conspire to find a place for her during the, relatively, good times during 1982, then she would get cut out again at each crisis. She was part of Novell's story in 1982, but it wasn't until after Ray Noorda's assumption of the presidency in 1983 that her position became permanent.
The printer was Novell's first released product. In 1981 the printer sold well, and the printer division was the most productive part of Novell's sales force.
But in the second half of the year, the printer business started having a problem that was later to plague the rest of the product line: there were a lot of returns. Most of the returns were DOA's (dead on arrival) -- the customer had opened the box, plugged in the printer, and it didn't work.
As the returns continued at a high level, finger-pointing started. At the center of the vortex was Production and its head, Joe Maroni. But he pointed, too, at Engineering, complaining that there were too many revisions to the product and too little pre-production testing. Engineering's reply was that all the revisions were to fix problems that Production said urgently needed to be fixed, or to add features that Sales said urgently needed to be included in the product.
Top management never broke the circle of fingerpointing, and poor product quality plagued Novell throughout its life as Novell Data Systems.
In the summer of 81, reality began to intrude on the Novell vision. It first showed in a classic way: lack of coordination. By early summer, Larry Edwards had built a sales force to sell the full Novell product line, but there was only the printer to sell. The terminal computer was barely released, and the terminal was still in development.
When trouble approaches a company, there are two places where it settles first. If the trouble is internal, the Chief Financial Officer will leave -- usually about six months before the bad news hits the streets. If the trouble is external, it will show up as turmoil in the sales force.
Sales people are quick on their feet. They're paid to be when they're paid on a commission plan. Novell salespeople were paid as is traditional for computer companies: a base plus commission. Commission is a compensation tool that sends a specific message to the receipient: it says, "the company is interested in what you produce today, not what you may produce some time in the future." If the person is producing immediate results, the company is interested in what the person is doing. If there are no immediate results, the person should be somewhere else.
Novell's sales were less than planned. Along with all the other problems low sales causes, it sends a message to the sales force that it's time for them to move on. Andy Olson, first aboard, was first to leave. He did so in May.
In summer of 81 Novell outgrew the Industrial Park building and moved half the company to Post Office Place -- twin buildings just being finished near the center of Orem, behind the post office. The engineering, design and manufacturing stayed behind. The adminstration, sales, customer support and marketing moved. (This complex, given up during the time of troubles, was later taken over by WordPerfect Corp and became the home of WordPerfect Magazine.)
With the move to Post Office Place came another signal of impending crisis. George announced that he would run the company from southern California. His wife didn't like Utah, so he was staying south and putting up an office there. Jack Davis was made General Manager to handle the Utah operations. It sounded to those of us in the trenches like George was feeling comfortable about the company and wanted to move on to other things.
This plan lasted for about two weeks, then George was back in town on his old schedule. The handoff had been announced, but it didn't happen. There was no new announcment to replace the old one, so it added to the confusion. The one clear message: things weren't going like they were supposed to go, and there would be more changes in the future.
By fall of 81, the printer was out and the terminal computer was out, but the terminal wasn't out and neither was the network. Products had been slow coming out, so sales were slower than expected.
This lead to the first shakeout. This first shakeout was a slight tremor, hardly more than a premonition. Roger felt it because he was in the Bay area, just about to sign on the dotted line of a lease for a Novell sales office he was going to head up in the Bay area. As he was about to sign, he got a phone call from Larry Edwards saying, "Don't. And fly back ASAP."
When he got back he had a meeting with Larry. Larry said that due to the slow sales, there were some tactical changes: Novell wasn't going to open the Bay Area sales office, and the terminal and the minicomputer were officially put on the shelf.
Putting the terminal and the minicomputer on the shelf had no great impact on Roger. He was a died-in-the-wool personal computer man, but the Bay Area sales office was to be his transition from Novell Customer Support to Sales. He had to wait for another opportunity. He didn't have to wait long. Two weeks later the head of the Dallas sales office announced he was leaving.
In spring of 82, another sign of impending crisis was the stream of managment consultants parading through George Canova's office door. When a manager doesn't have an answer, he or she seeks one out, and a common source for answers is a consultant. The sign of trouble is when there are a lot of consultants.
The consultant activity goes up when the manager isn't hearing answers he or she likes: answers that will solve the problem at hand without creating a lot of organizational pain. If the first consultant doesn't have a good, painless answer, then it's time to find another consultant. If the consultant stream is growing and the turnover is high, then trouble is afoot.
The second sign of trouble is when the consultants aren't experts in the problem area. George's most frequent consultants, for instance, were old business associates of his with extensive business experience, but with little experience in this new technology or market area that was the personal computer industry. These consultants didn't have any more specific experiences in the personal computer area than George did. They could offer general truisms, but not "the word on the street" -- the specifics that made this market different from any other. This showed up in the questions the consultants asked of others in the organization as they did their research, and it was soon clear that they didn't know much.
It was also clear that some carried great influence. One was Jim Walker, the person who recommened the infamous COBOL accounting package.
Just before leaving for Dallas, Roger listened to Jack give a speech that sent shivers up and down his spine: it was a severe attack of deja vu!
Another excerpt from Roger's diary.
"Jack communicated a lot, but it wasn't his habit to give group pep talks. But this particular meeting sounded like that was what it was to be. All the employees at the Post Office Place building gathered in the central meeting room and Jack stood there in front of us.
"A quick summary of the speech was as follows:
1) We aren't in trouble
2) We need to cut expenses.
3) We can't sacrifice service while we cut expenses.
"The suggestions he gave were: don't use Federal Express when mail will do, and use customers' 800 numbers whenever you call.
"I had the impluse to jump and say, 'And don't use shipping boxes when envelopes will do.' then turn to him and say, 'That's what I told my people.'
"The deja vu came from the fact that I had given the same speech to my staff just nine months previous when I was running Computerland, and these were the circumstances that lead me to make my speach: I had just completed a series of financial projections -- using Visicalc for my first serious spreadsheet project. The results were discouraging: using a scaled down projection of sales -- based on the real sales I'd generated for the six months previously -- and a continuation of expenses as they had been for those same six months, I couldn't see a profit.
"For those six months I'd done everything I could think of to boost sales, so I didn't see any way to make any dramatic improvments in sales volume. I still had to buy product to sell, so I couldn't reduce inventory expense much. I didn't feel like there was any 'excess baggage' in my staff, so I couldn't see how to make many cuts there. That left administrative expenses, so I called my staff together one day and gave them "the speech": 'We need to cut expenses but not service, so don't make long distance calls you don't need to -- but if you need to make a long distance call to generate business why be sure you do! -- and lots of other similarly contradictory directives.
"The speech was, in effect, an admission on my part that no action I could think of taking would save my Computerland store situation. Only a change in the business climate Computerland operated in -- something that spontaneously picked up sales, fattened margins or lowered costs -- could save the day.
The day I made the speech I started looking for a buyer.
"Now, a year later, Jack and George were facing a similar situation at Novell. They were nearing the end of their startup money. Sales were below plan, but without a sales force they would never reach plan, so the Sales force couldn't be cut. (Much more, that is. In fact, the sales force was cutting itself very effectively as salespeople who weren't making any money left of their own will.)
"Development couldn't be cut much because half the products weren't out yet and the other half were still buggy or not quite what the market needed.
"Manufacturing couldn't be cut much because products had to be produced to sell and there were a lot of returned goods that needed to be processed and repaired.
"In short, Jack finally came to the same course of action I did: make a speech about cutting administrative overhead because all the other quick fixes were being done to the limit that they could be done. Then plan for some radical change.
"As I heard the speech, I knew it signaled a change of heart in Jack, and probably George. Their view of Novell's potential had changed, and some radical change in the company was forthcoming. What form it would take was still a mystery, but it wasn't likely to be pleasant for the founders."
While storm clouds were growing over Novell all through 1981, there was lots of sunshine too. One of the bright spots was the evolution and development of the local area network.
Novell's product line -- intended and real -- was never static, it evolved from Day One. And it evolved in logical, internally self-consistent fashions. Jack started the company on his vision of a multi-CPU minicomputer -- something that would break out of the limited-applicable-size constraint that minicomputers of the seventies offered. This minicomputer would need terminals and printers -- something Jack was particularly experienced at selling, so why not build these too? First, in fact, to finance the development costs entailed in designing the minicomputer. So, in the summer of 80 the potential product line consisted of a minicomputer to be developed and terminals and printers that would be built first to provide money for the mini's development.
By fall of 80, it was clear that Safeguard would be doing the bulk of the financing. They wanted something their Business Systems dealers could sell. George said, "These people certainly can't sell a multi-user mini. This means a standalone microcomputer... no problem, we'll convert our terminal into one. Doing so is an industry trend right now." Thus, the terminal computer was born and added to the product line as another way of bringing in revenue to support the minicomputer development.
But by spring of 81, it was clear that the mini would be a long time coming, and that the terminal computer needed some help. The terminal it had been based on was a feature-rich design, intended for the Burroughs/Univac marketplace, where big screens and weighty terminals are traditional. It was competing in a microcomputer world accustomed to simple, small-screened, light-weight ASCII terminals. This made the terminal computer look big, feature-rich, slow (because of it's graphics capabilities) and expensive. One way to speed its performance so it didn't look slow was to add a hard disk, but this would increase the expense even more. One way to cut the expense of adding the hard disk was to split it among several terminal computers -- make a network. This was evolution that produced a "need" for a network product by the spring of 81 (more on this later).
In the spring of 81 Jack was still pursuing his Novellwriter dream. Since none of his commercial contacts had born fruit, he decide to end-run the same way he had with Autogen. He contracted with some BYU grad students to develop a word processor.
Just days before these BYU programmers drove up to the sparsely populated parking lots at Industrial Park Drive, Novell's top management conducted a brainstorming session: sales of the terminal computer were growing too slowly and something needed to be done.
The managers assessed what they had created:
o It was a pricey unit.
o It was a graphics terminal with a big screen and programmable function keys.
o It was one of the first CP/M computers to come with a hard disk as standard equipment.
One of the question asked was:
"Can we reduce the cost?"
One of the answers was:
"Yes, by sharing that expensive hard disk between several terminal computer units."
To do this, a local area network would have to be developed.
Among those top managers, the LAN idea had a lot of supporters:
Jack was aware that two startup companies, Corvus and Nestar, were selling personal computer-connecting LANs, so he could see a market developing.
For Jack and Sherrill the LAN was a way to keep the 68000 chip as part of the product line, even if the multi-CPU minicomputer product was delayed.
For Dennis, the LAN was a different way to transcend the "limited range of users" problem that afflicted all traditional minicomputer designs. Instead of adding CPU power by putting additional cards into the central box, the CPU power would be added with each terminal added to the system.
For Larry Edwards, it meant something to sell sooner rather than later.
The idea had a lot of top managers going for it, and when the programmers from BYU arrived, the word processor was never mentioned. Comdex was coming!
When the temporary BYU programmers, Drew Majors, Dale Neibar, Kyle Powell and Mark Hurst arrived, they reported to Phil Long, the head of Software at Novell. Phil and his regular staff were working long and hard on the Safeguard project -- making the terminal computer a suitable product for Safeguard Business Systems to offer it's One Writer dealer/customer system.
Phil introduced the expression "creeping feature creature" to Novell's lexicon. When new ideas came up, he was often the one to ask, "Sure it's nice, but do we really need it to sell product?"
When the LAN project came to his doorstep, it was just another feature, and he didn't have the resource to devote to yet another feature-adding project. So Jack borrowed his new temporary talent, the BYU boys, to get a LAN demonstration built for Comdex 81.
Dale and Kyle started working on the terminal computer side. Their task was to pry the CP/M operating system apart far enough so they could get disk requests to flow out over a wire to a remote file server rather than go to an internal disk controller. The programmers needed something to test their work on, so Drew took another of the terminal computers and started developing a testbed file server.
Dale relates, "We'd been in this position once before. A year earlier we were working as part of a team project and the other part of the team failed to deliver. We had no way of proving our part worked, so we didn't get paid. This time we weren't taking any chances."
Fate proved the wisdom of this strategy. The fellow that was developing the 68000 file server side had a motorcycle accident. In a flurry of long hours and last-minute hustle, it was the Z-80-based testbed file server, stuck over the top of the 68000 board, and using RS-232 connections, that was shown at Comdex.
This demonstration system was an example of "blue smoke and mirrors" -- an expression Andy Olson liked to use. Novell used blue smoke and mirrors to make a point at that Comdex. That first system was never built to be sold, it was built solely to show what Novell was going to be up to in the coming months.
This practice is controversial: to some it is "fooling the public" because the thing being demonstrated may or may not come to fruition.
The value of showing off this "may become a product" is that it's risk-reducing to the vendors and consumers of products. It's a step in test-marketing a product. It helps vendors find out how valuable potential customers perceive this creation to be, and what adjustments need to be made to make it even more valuable.
Not bringing prototypes to a trade show raises the cost of product introductions. It forces companies to spend more money before demand is proven. We end up with more Edsel cars and Premier cigarettes, where a lot of money is spent making a product that won't sell. Or we end up with more America-losing-the-FAX market situations -- where a market isn't tapped because the demand looks too risky for the amount of up-front money that must be committed.
Blue smoke and mirrors -- prototyping -- is a powerful cost-reduction and feature-tweaking tool for marketers. But like any other powerful tool, it can be abused as well as used.
The LAN demonstration at Comdex was a hit; lots of people were interested. George and Jack had pulled off another coup. They'd come to Comdex 80 with nothing but promises to be a big company, now at 81 they were showing off cutting-edge products: the printer, the personal computer with a hard disk, and the LAN.
It didn't take Jack long to sense that this LAN concept was generating more interest than he had expected. At a lull in the bustle of the show, Drew pulled Jack aside, "You know, Jack. This LAN business is something we programmers have been thinking about between us before now. And what you've been describing here at the show is close to some of the ideas we've developed." Drew and Jack headed home from Comdex early, and within a week the temporary programmers who came to build a word processor -- and who were later going to become SuperSet -- formally became the core of the LAN development team.
One of the first issues the proto-Superset team faced was the question of how to split the disk-controlling duties between the personal computer and the file server. The simplest and quickest approach was to let the personal computer do most of the work -- the most important piece of that work being deciding where on the disk to store the data. This is the "disk server" approach. The disadvantage of having the personal computers deciding where to store data is that the file server can't coordinate between the various computers easily.
The quickest and easiest way to solve the coordination problem is to not have any coordination -- each personal computer gets allocated a fixed portion of the hard disk and there is no overlap. This is called disk partitioning and it's the simplest of the disk server approaches.
Disk server: quick to develop; simple in concept. Larry Edwards and the sales force loved it. They could say, "Yes, we have a LAN." and be selling it within months. The benefits were so compelling that most of Novell's early LAN operating system competitors were disk server-based.
The alternative is to give more duties to the file server -- let the file server decide where to store data on the disk and let the file server handle coordination between the various personal computers' requests. This is the "file server" concept. It requires the file server operating system to do more, and it requires more modification of the personal computer operating system. The benefits are more coordination between the personal computers and more efficient use of the disk storage -- partitions aren't required.
There is one other serendipitous difference between the two: file service requires more horsepower from the processor than disk service does. Novell's plan to use the 16-bit 68000 for the file server -- while most of the competitive units were based on the 8-bit Z-80 -- made file service easier for the Novell team to develop than it would be for competitive teams.
Thoughts of disk service versus file service were churning through the minds of the Superset in those weeks after Comdex 81. The more they thought, they more they wanted to develop file service, and not just file service but file service with system fault tolerance. Doing this would be difficult in the face of Sales demands for "something I can sell today", but as it turned out, other events in the company were going to force themselves into the limelight and give the team the time to follow their dream.
One of those events was the firing of Jack Davis.
Jack and Drew flew back from Comdex flush with success. It looked like they were finally on the track of something the market really needed from Novell, and it looked like Drew and the other programmers had put enough thought into this concept that they could produce superior results.
But the news waiting in Utah wasn't good. The first shoe had finally fallen, and Jack was no longer a part of Novell. The hard times "the speech" had foretold were happening, and the first casualty was Jack.
1981 was the year in which Novell defined itself. It transformed from an idea into a company: a company based in Utah: a company that would sell printers, personal computers and local area networks. It was also a year in which many ideas were cast off: it would not be a company which sold minicomputers or Sperry-compatible terminals, it would not be a company of which Jack Davis was a part.
The challenge that Novell faced coming into 1982 was that what had been defined as Novell, thus far, was not profitable: it was still consuming more resources than it was creating. With the firing of Jack Davis, the writing was on the wall, the company would have to become profitable soon.
Larry Edwards had been able to assemble an experienced sales and support staff very quickly. Novell never lacked for salespeople in 1981.
The promotion and marketing efforts of Jack Davis border on legendary. A company that didn't even exist in 1980 was getting media attention at Comdex 80, and by Comdex 81 was promoting like a well-established minicomputer company.
Rusty and Phil had created a high-end, innovative personal computer that any ex-IBM mainframe or ex-Burroughs terminal user would feel comfortable with.
Finally, the company as a whole had recognized that local area networks were going to be important to the company's future. The LAN was never a "skunk works" project that had to be developed in secret because part of the company was dead set against it. Everyone knew the LAN would be important. It just wasn't getting there soon enough.
Jack and George had assembled a highly professional team, but they couldn't supply the vision to keep them all pulling in the same direction. These people, including George and Jack, brought with them old habits and old experiences that were appropriate for the minicomputer world. Some of these experiences were appropriate for the personal computer world, but some were not. Mistakes were made, many of them costly.
The evolution of product had produced results that were different than what the mainstream of personal computer users were expecting. The Novell system was high-end at a time when the market was still more interested in low-end solutions. For instance, it offered graphics, but no way for the average user to use them. The average user could get more graphics use out of an Apple II because there were software packages available for the Apple II that utilized the graphics. These features added cost, but didn't add customers.
Jack was fast at assembling a plan, but he was poor at checking for holes in his logic; George didn't have the background to fill in those holes. Ray Noorda points out, for instance, that manufacturing electronic hardware in Utah was a fundamentally risky proposition in the early eighties. Jack wasn't oblivious to this. He had helped Televideo import their terminals, and every day he drove past the abandoned ADDS terminal factory that Harry Armstong had once worked in. But on the other hand, even in the early eighties, there was lots happening in electronics in Utah, and George's background was in manufacturing. If anyone could make domestic manufacturing happen, it was George, and he was willing to try.
One of the most costly mistakes was the inability to resolve the quality issue. Finding fault is only important in as much as it's a step towards resolving the problem, but top managment never finished resolving the quality problem, so the organization wallowed in the fault-finding stage, and it's customers stayed away in droves.
So, by the end of 1981 Novell already had a checkered past. Some of it's efforts were hits and others were misses. It would take another year, 1982, to sort out its future.