In Memorium of Ray Noorda
Ray Noorda passed away on October 9th, after suffering a long battle with Alzheimer's. Unfortunately, most of the obituaries I've read seem to have taken a "by-the-way" approach to his passing, and almost all of them have utterly failed to convey the tremendous impact he had on our industry. At a very minimum, this is a lost opportunity to revisit some of the lessons he taught.
According to the official bio, Ray Noorda was born in Ogden, Utah in 1924, the son of Dutch immigrants. His early life was spent in the depression, working odd-jobs to make ends meet. During WWII, Mr. Noorda joined the Navy as a radar technician, and then subsequently obtained a degree in engineering from the University of Utah in 1949. From there, he went to work as an engineer for GE, where he remained for 21 years in various roles, including stints in marketing and management. After leaving GE, Mr. Noorda worked for a variety of California companies, plying his technical and management skills to become a successful turnaround artist for troubled technology firms.
In mid-1983, Mr. Noorda joined Novell as President and CEO. At the time, the company was one of a handful of providers of hardware-based PC networking equipment, and was struggling to make money in a nascent and ill-defined market. But Novell had a secret weapon—a handful of student contractors were secretly developing a software-based networking platform, which Mr. Noorda pushed into packaged form by the end of the year, and which subsequently became known as NetWare. Although the initial product releases didn't exactly set the industry on fire, it was eventually ported to common PC hardware, and within a few years NetWare 286 was positioned to the lead the PC networking industry into a new era.
People don't really remember, but prior to NetWare, "PC networking" mostly consisted of buying a custom box that was little more than a specialty hard drive with multiple connectors, which you would attach your PCs to. Moreover, most of the computing industry as a whole was still pushing vertical solutions, especially mid-range computers and multi-user systems, and that was the model people understood for early PC networking as well. A software-based solution using PC hardware was considered to be a radical proposition at the time, and many people dismissed the proposition outright.
As we all know by now though, software can be compelling. Not only is it more flexible and therefore ultimately more powerful than fixed-in-time hardware architectures, software's independence from hardware lock-in also put the buyer in charge of their own future. You could build your own systems to suit your specific needs, usually for less money than you could buy an ill-fitting vertical system, and without having to commit your entire enterprise to a vendor's architecture. If a vendor's system didn't work as a NetWare server, you could just swap it out for another one and still keep all your data and other assets intact, a promise that was unmatched in any other sector of the industry at the time. This seems obvious today, but it was practically unheard of back then.
In fact, given that a large part of the early PC success story came from networks that were pieced together beneath the corporate IT radar, Novell is arguably responsible for much of the PC industry's early success, since it was the principle technology that actually allowed people to move away from vertical architectures towards distributed PC computing architectures en masse. Sure, Novell benefited from the PC boom, but the inverse is also true—without Novell and specifically Ray Noorda, the PC boom itself would have almost certainly evolved down a different path.
Providing the software catalyst for this change was significant, but it was not enough to move the entire industry, and Mr. Noorda's true genius shows through in the secondary work that he undertook to ensure success. For example, Mr. Noorda recognized early on that one of the hindrances for wide-scale adoption was the relative high price of PC networking hardware, such as Ethernet adapters. So, the company borrowed a reference design for a cheap Ethernet card from National Semiconductor, and had it manufactured for sale at a low cost. This simple act drove down the cost of building a NetWare network and helped drive sales of NetWare, but it also had the secondary effect of making Novell a leader in the Ethernet hardware space, as their low-cost cards were picked up for use by other markets as well. As a present-day observation of the significance of this simple act, many of the network cards that are sold today still claim to be "NE2000 compatible" so that they will continue to work with the broadest set of platforms.
Mr. Noorda is also widely credited for developing "the channel" distribution system that our industry currently depends on. Software certification for technology "engineers" is another marketing technique that was pioneered by Novell under Mr. Noorda's leadership, and is still heavily used by Novell's competitors today. Mr. Noorda also enunciated and advanced the concept of "coopetition" whereby competitors could cooperate long enough to build the industry, while still competing for customer dollars in the larger market. All of this contributed greatly to the success of Novell and NetWare, and within a decade of release, the company was doing a billion dollars in business per quarter, with NetWare retaining 70% of a booming PC network market.
Novell did not reach their pinnacle without notice or competition, and NetWare was attacked from multiple angles throughout the latter 1980s and early 1990s. Vendors such as LANtastic tried to undercut them at the low-end, while companies like IBM, 3Com and Microsoft tried to squeeze them out at the high-end. But as the software market heated up, Novell made sure that NetWare was always the most functional, with the fastest performance and the most features (including technologies such as fault-tolerance, cross-platform connectivity, and eventually directory services), as well as the broadest reseller base and a hoard of fanatically loyal customers, thereby retaining the majority market position. Before it was over, IBM was selling NetWare in a blue box, 3Com had effectively abandoned the networking software market, and Microsoft had to bundle their networking code into Windows NT to get any significant market share.
In short, Mr. Noorda retained Novell's leadership position throughout the first decade of PC networking by always ensuring that the company was driving the industry—first with the move from hardware to software, and then later with service and features. But Novell was not prepared for the market's transition to networking as a commodity service, especially one with a greater emphasis on cost than value. Simply put, general-purpose operating systems from Microsoft, Apple and *NIX vendors began including free networking services in their base OS offerings, and the need for a dedicated "network operating system" began to diminish. By the late-1990s, the end of the line for NetWare as a mainstream commercial product had become plain to most people.
To Mr. Noorda's credit, he tried to react to this shift by realigning Novell at the end of his tenure. For example, Mr. Noorda oversaw the initial development of UnixWare, a general-purpose UNIX operating system for Intel-based PC hardware that was quite stunning for its time, then acquired Unix Systems Laboratory (USL) and all rights to UNIX in 1993, and then settled long-standing licensing disputes between USL and BSDI so as to strengthen the position of UNIX as a whole. As one of his last acts in office, Mr. Noorda also acquired GroupWise, WordPerfect, Quattro Pro and a variety of other productivity applications in June 1994 in an effort to diminish Microsoft's budding dominance in that part of the market as well. Mr. Noorda arguably made some mistakes in his zeal to attack Microsoft, such as pursuing Digital Research's DR-DOS as a lawsuit vehicle against Microsoft, and sinking development into GEM as an alternative graphical environment to Windows, but in terms of overall strategy, he was demonstrably correct to pursue diversification, and many of his initial steps have since been mimicked by other vendors.
But by 1993, Mr. Noorda had already begun to suffer early symptoms of Alzheimer's, and his days at the company were already numbered, even while he was driving the realignment. According to historical SEC filings, Mr. Noorda first resigned as President and CEO in April 1994, and then resigned as Chairman of the Board in August 1994, and then was forced to leave the board entirely in November 1994. Over the next three years, subsequent management teams at Novell abandoned Mr. Noorda's diversification projects (first by selling off UnixWare to SCO in 1995, then by selling off WordPerfect to Corel in 1996), leaving the company with just a handful of significant products, a declining share of a collapsing NOS market, and few significant opportunities for growth. Although the company has made recent moves to correct some of these mistakes (mostly by replaying Mr. Noorda's previous strategies, most notably by acquiring SUSE Linux in 2003, and also by porting the bulk of their networking products and technology to run on Linux), it's reasonable to believe that the intervening ten years of pain and misery for the company could have been avoided if they had simply stuck with Mr. Noorda's original vision.
After his exile from Novell, Mr. Noorda directed his efforts into the venture capital firm called Canopy Group that he had started in 1992 with the intention of investing in Utah technology firms. One of their better-known investments was in a company known as Caldera Systems, which was launched with the purpose of building a commercial version of Linux (according to Ransom Love, Caldera had also been started as a back-room project at Novell, but was killed by the post-Noorda management team, and then resurrected as an independent venture by Noorda at Canopy). Caldera initially met with some success in the industry—Network Computing awarded Caldera OpenLinux 1.3 the NOS of the year award in 1999, beating out NetWare and Windows NT alike—and at the time it appeared that Mr. Noorda had managed to replay his winning formula, driving the development of a cheaper and better technology, and changing the industry along the way. But eventually this effort would also fall apart.
In August of 2000, Caldera acquired most of The Santa Cruz Operations' UNIX assets, using a cash settlement from Microsoft as a result of the old DR-DOS lawsuit (Caldera had acquired DR-DOS from Novell for an embedded systems sideline business they were developing). As part of this purchase, Caldera acquired SCO OpenServer, as well as the Novell UnixWare platform which had itself been acquired by SCO in 1995. In 2001, Ransom Love left the company, and the next year Caldera renamed itself to The SCO Group. In 2003, the company began to launch a series of lawsuits claiming license violations in Linux against SCO intellectual property. Caldera Linux products have since been abandoned.
In 2004, Mr. Noorda's daughter led an ouster of Canopy Group CEO Ralph Yarrow (a former graphic artist for the Noorda family's philanthropic organization who was promoted to CEO of the VC firm by Mr. Noorda) and two other associates, claiming they had taken advantage of Mr. Noorda's declining mental state to pilfer $20 million from Canopy Group. In a last-minute settlement, the Canopy Group traded their shares of The SCO Group to Mr. Yarrow and his associates in exchange for their interest in the Canopy Group, effectively buying him out of the venture firm. A few days later, Mr. Noorda's daughter committed suicide in her family home. Mr. Yarrow still serves as Chairman of the Board of the SCO Group, but the company no longer bears much of a resemblance to the former SCO or to the Caldera project that was funded by Mr. Noorda.
Recently things had begun to improve for Mr. Noorda. Most notably, his two sons had taken board positions in the Canopy Group, which has since returned to its mission of investing in Utah startups. Meanwhile Linux and UNIX in general have come into their own as viable broad-scale PC-based alternatives to Microsoft Windows largely due to Mr. Noorda's early and influential efforts with UnixWare and Linux. Meanwhile, Sun's acquisition of StarOffice and Google's development of web-based office tools as vehicles for weakening Microsoft's hold on productivity applications were also preceded by Mr. Noorda's maneuvers at Novell a decade prior.
As for the rest of us, we all work in an industry that was largely defined by Mr. Noorda during the early years of Novell. For these and other reasons, we should all be grateful to his contributions.