"I came here to drain a swamp, but I spend all my time wrestling
alligators."
--Anonymous
In 1982, Engineering was tied up wrestling the enormous alligators of quality problems, personal computer problems, and getting the LAN delivered. The continual battering had left them quite conservative and focused. When the IBM PC was introduced, few engineers showed any interest. They had no time to see that the IBM PC was the key to draining the swamp.
Superset, on the other hand, was quite interested. They were independents and their contribution was software, to which they still felt some proprietary attachments. They could see Novell's problems as clearly as anyone, and if Novell sunk beneath the waves, they wanted to have another market for their work.
All through 1982 they actively explored alternatives -- some were marketing alternatives, such as discretely inquiring as to who else might be interested in network operating system technology, and selling a game they designed to performance test the network (this was the origin of Snipes).
They also explored technology alternatives. One of the technology alternatives was the IBM PC. Drew got one of the first PC's in Utah valley. He could see that it would be easy to network as a workstation. It had an open bus structure and a well documented BIOS. These would make designing a network board easy to do. So in 1982 and 1983 Drew campaigned, with Craig's help, to include the PC as one of the kinds of workstations Novell would support. It was fine as a workstation, but as a file server it was marginal: it had no standard hard disk.
In 1983 the PCXT was introduced. The PCXT was a PC with a standard hard disk installed. Drew saw that the hard disk obstacle was removed. The PC-compatible file server was now technologically practical, and supporting PC-compatibles as file servers became a marketing question. There were some serious questions about doing so: the PCXT was "underpowered" compared to the proprietary 68B file server, and the 8088 CPU couldn't process as fast as the 68000. The PCXT supported smaller disks -- 5MB and 10MB compared to the 40MB available on the 68B. Would such a file server sell? Finally, even if it would sell, would it cut into the much more profitable business of selling 68B file servers? 68B file servers sold for a few thousand dollars. An operating system for a PC-compatible would sell for a few hundred dollars.
Craig and Drew argued that it would not only sell, it would sell much better than the 68Bs because the PCXT was an open system. There would be dozens of companies and thousands of people trying to sell PC-compatible file servers with dozens of kinds of networking boards installed. If Novell could service that market with a "glue product" it would be a much larger market than the market of selling just 68B file servers with Novell boards.
But it was a tough choice. It went against the grain of what Engineering perceived their mission to be, a circuits designing engineer isn't going to have much to do in a company devoted to producing floppy disks and manuals. That engineer is going to perceive that he or she can add a lot more value to the company by redesigning the file server or designing a new network interface card. (In the end, the circuit designers adapted to the "glue product" concept, and proved useful by designing drivers for NIC board and by dealing with the many kinds of disk drives and disk drive controllers that came to be attached to the PC-compatibles.)
The path to Novell's famous open systems approach to marketing Netware was not an obvious one in 1983. It took a lot of inspired salesmanship both within and without Novell.
When viewed from the outside, 1983 was a year of no change for Novell. Novell hadn't grown in people, the people hadn't changed much, the LAN product hadn't changed much, the sales were only 3.5 million. There was absolutely no sign that a billion dollar company was in the making.
The only surprising news was that Novell was still alive at all, and that the chaos of the Time of Six Presidents had ended abruptly with Ray's arrival.
But in 1983 the foundation was built: the key elements of Netware were resolved technically, and now the battle would move to marketing this wonderful technology. The task ahead would be to prove that Netware wasn't just another "over-engineered product" -- a product that had lots of nice features, but features that no one needed.
In 1983 the team was forged to prove that Netware was a worthy product, and the goal for that team was defined: 1984 would be, in the eyes of Novell's founders, the Year of the LAN.