The Love of Linux Chapter 3: The Birth of a Company
Bryan, being a systems engineer, did not want to continue with just the interfaces, so he spoke with Ray about funding the Linux effort outside of Novell. Bob wanted me to stay on and work with Novell to get the work transitioned to a new group formed in California for the client. I had spent some time with the Lan Workplace team in California. This team had lobbied very heavily to get our team assigned permanently to them. Lan Workplace made their business selling TCP/IP stacks to the different client systems needing to connect to the Internet. Again, this was before Microsoft included TCP/IP with every operating system. They saw everything we did as a direct threat to their business even though it was the best thing Novell could have done for the server business that brought in nearly 90% of the revenues at the time. To appease Bob, they had told him they wanted our team because they needed a browser.
Bob Frankenburg had arranged to demonstrate the higher level interfaces at his Fall Comdex keynote and had asked me to participate. After the rehearsal, I had a few moments to talk with Bob alone. I told Bob that if he did not continue the efforts we had begun on with the HTTP server, WordPerfect applications, as well as the enhancements to the client, the work would not see the light of day. Novell decided to just work on the client so I resigned and joined Bryan as one of the co-founders of Caldera, Inc. Within three to four months the project inside Novell was dead. Novell had all the right technology, the market share, the cash, the right engineering talent, but without Ray and some of those who started the effort, they had limited vision. Most of the visionary talent around Ray in the early days had left. The original team had more of the entrepreneurial spirit. Much of the management that was left was middle management that did not understand how to take risks and provide leadership for the company, let alone the industry. Some were good managers but they had not had at the time, the experience of growing the company and did not know how to take appropriate risks even though some went on to do great things. Nearly every business function requires those who pursue the discipline to become go efficient managers in order to be promoted. Often managing resources means minimizing risk. When men and woman are promoted from line management into executive positions, they cross a chasm. Executive management, while being able to run a business efficiently, they must also be willing to take important strategic risks. They must be entrepreneurial or be able to think strategically and separate themselves from the operations. Many very talented individuals cannot make this switch and businesses are managed into the ground or key strategic opportunities are missed in favor of pressing operational demands. In addition, I learned from watching Bob and some of the managers at Novell that even very talented men can have limited success when they surround themselves by the individuals who tell them what they think they want to hear or who are not willing to take appropriate risks even for the right reasons. I listen to Bob speak at Utah Valley Community College recently and he mentioned that one of the biggest challenges he faced at Novell was peoples’ unwillingness to tell him the bad news. I will describe in more detail how Caldera went about creating its Linux distribution, but it is interesting to note that Novell rejected Linux at this time and Caldera was formed. Caldera acquired LST in Germany to provide a business quality Linux to the market. After going public and purchasing SCO, Caldera, to cut costs and gain developer support consolidated its Linux efforts with SuSE to form UnitedLinux. The core of Caldera’s Linux engineering team went to work for and began to develop SuSe Linux. Then, Novell purchased SuSe for $200 million and essentially purchased back what they had germinated. Many things in life are cyclical and the world is a very small place. Do not mistreat those you work with or cheat your customers or partners. Life is just too short and the world too small.
When Caldera began, the “distribution” business of Linux was in its infancy. Slackware was probably the first company to begin distributing all of the components that make up a complete Linux operating system environment and distribute them on a CDROM, hence the term, “distribution.” Prior to Slackware, all the components were simply downloaded from the Internet. It was not Caldera’s desire to go into the “distribution” business. At the time Slackware was by far the dominant player. Yigdrasil was the up and comer. Caldera wanted to continue what we had started, add enough value above the operating system to create an alternative desktop in the industry. Since we did not want to go into the packaging business, we decided to partner with Redhat. We chose Redhat because they had the rudiments of a packaging system needed to handle the operating system and layers above the operating system. Bryan and our engineers went to work on Red Hat Package Management or RPM. Bryan and I flew back to Connecticut and met with Bob Young. Bob was the manager and partial owner of the ACC Bookstore. He had acquired the marketing rights to the Redhat distribution. The distribution was produced by a couple of students at North Carolina University. We made an agreement with Bob to produce a distribution only once in every 6–12 months. At the time, Redhat was producing a new distribution every 2–3 months. We knew we needed to slow the process down to attract commercial developers. We gave the architecture for RPM to Bob and paid Redhat to make the enhancements to RPM. We then went to work on the upper layers.
Within 6 months, a team of 5, Bryan Sparks, Ransom Love, Jim Freeman, Ron Holt, and Nick Wells produced our first commercial product, the Caldera Network Desktop. It contained the first graphical desktop for Linux ever produced, a product we licensed, ported and modified, from Visix. The Network Desktop sold over 12,000 copies and was priced at $99.00. We thought we had failed. In reality, we had probably sold a copy to every Linux enthusiast there was at the time. Not to mention, the average selling price of a Linux distribution was between $8–12. The Caldera Network Desktop was a phenomenal success. It put Redhat on the International stage and lifted them above all other contenders. Those who had purchased the Network Desktop wanted all of the latest changes to the kernel. Bob Young wanted to appease their thirst, but he had a contract with us. Caldera realized that we needed to be able to control the bits and bights to ensure a business quality product. We parted ways. To their credit, Redhat knew who was buying Linux at the time; the hacker/developer and they wanted timely updates to the code. Caldera was still pursuing its dream to reach the commercial buyer.
In the process of developing the Network Desktop, Sebastian Hetzie, a Linux pioneer in Germany, had partnered with us to form the European arm of Caldera. Sebastian introduced us to a small group of engineers who had created their own distribution in Germany called LST. Ralf Flaxa and Stephan Probst had helped us successfully improve the installation of the Network Desktop. When Caldera parted ways with Redhat, we negotiated with Ralf and Stephan to purchase LST. Caldera officially started to produce their own distribution targeted at commercial users. Our only problem was there were very few commercial users at the time. The open source model was not understood by us or anyone at the time. It was just emerging. The industry in some ways, was somewhat caught up in the hype and the business acceptance of Linux was still coming. For example, a survey published on Slash Dot just a month or two before Redhat went public revealed that 70% of Redhat’s customers were between the ages of 12 and 25, hardly those who were representative of commercial purchasers, but they were a strong, very vocal customer base who believed they were part of a revolution.
Caldera was still focused on delivering a business solution. We knew we had to get business applications. At the time, our only hope to get more applications for the desktop was using DOS. A group of engineers had developed the ability to run DOS applications in a window on Linux. If we could get access to DOS, we felt we could help the effort along. In July of 1997, we acquired from Novell, the DR DOS or Novell DOS asset. We wanted DOS to provide further integration with Windows on the desktop and we knew at the time that even if we could build better products than Microsoft, Microsoft had a strangle hold on the hardware manufactures in the industry and we would not be able to get the hardware companies to bundle our software. Novell would not sell Caldera the DOS asset unless we filed an anti-trust suit against Microsoft. Novell had done the work on the suit and assembled a very good legal team, but either management or the board where afraid to attack Microsoft head on in a suit. Novell had found some of the best attorneys in the industry that would take the case on consignment. To break the strangle hold on the OEMs, Caldera felt we needed to fight Microsoft in a court of law as well as build better products. While we were correct in our assumptions, this decision proved to cast a shadow over the success of our company. Many companies did not want to work with us because of our pending suit with Microsoft. I realized first hand just how many of the largest companies in our industry lived in fear of the wrath of Redmond and still do. Novell’s board would not bring the lawsuit against Microsoft themselves for fear of retaliation, but Novell’s corporate attorney, David Bradford was certain of its outcome and wanted Novell to share in the proceeds. We knew it was a matter of discussion at Novell, as two individuals who associated with the David approached Ray separately about funding a company for the sole purpose of bringing the lawsuit against Microsoft.
Novell’s attorney, David Bradford proposed a way to “obfuscate” Novell’s participation in the lawsuit by creating two agreements. One was a Novell DOS asset purchase, and the other a licensing agreement. Novell agreed to license us many of their technologies, like NetWare for UNIX, GroupWise, Personal NetWare, and several others. They then set a royalty limit that allowed them to recognize a 20% royalty on any revenues above 10 million. The 10 million dollar amount was determined to be the natural growth of Caldera’s revenues over the three years the agreement was in place. We felt the deal would be very beneficial to not only Caldera, but the entire Linux industry by bringing significant commercial applications to Linux. However, the Novell management team was not behind this agreement as they did not follow through with many of the license commitments. From our perspective, it appeared that Novell’s board or management team simply wanted to sell DOS. Some of the board members may have liked the idea of the proceeds from the suit, but they did not want the political or even business ramifications of entering into a suit with Microsoft. Since Caldera was interested in DOS and David Bradford knew that Redmond would not intimidate Ray Noorda, he tied the sale of DOS to the suit. The only way Caldera could get access to DOS and the applications was to sue MicroSoft. Because David felt the suit was such a sure thing, he did not want to give up on it so my guess is that David then sold the concept to the board but failed to follow through and convince the management team. The technologies offered would have significantly helped move the Linux effort forward, but we were unable to get Novell to deliver Groupwise, Personal Netware and the other technologies committed to by contract. David Bradford and potentially other members of the board had what they wanted, their hooks into the potential Microsoft suit proceeds, but they did not bother to get management’s interest or buy in to push through the technology licenses.