The Love of Linux Chapter 2: The Business Plan
Nothing happens by chance.
At the time, Microsoft was refusing to expose Novell’s network services as they were trying to force users into purchasing NT. Microsoft was in many instances, giving NT away with those who utilized their desktop to try to take market share away from Novell. Bryan, Rob and I felt that the same could be done with the desktop. If you could develop the desktop utilizing a different paradigm that did not cost millions of dollars, you could focus a small group of engineers on the key user interfaces and leverage all the services provided by the server. Linux provided the ideal method, open source. With the operating system being developed by the community, the traditional costs of developing the total operating system could be greatly reduced so that the desktop could be almost given away with the server. Rich backend services could be provided by Netware and if nothing else, we would force Microsoft to expose the NetWare services in their desktop to compete. Besides, there was no Graphical User Interface at all on Linux at the time. To get Linux mainstream, we needed much richer tools and user interfaces.
The best option would be for Novell to do this, but we did not feel that there would be any interest. So, the original idea was to start a company outside Novell. We began writing a business plan. Rob Hicks mentioned our ideas to Darl McBride who was the Vice President over the Novell Embedded System Technology, or NEST group. In a meeting with Ray Noorda, Darl mentioned our ideas to him. Ray immediately called Rob in and spoke with him. After several long discussions, Ray decided to sponsor our ideas inside Novell. We met with all of the Vice Presidents and they were all very much against the idea. They kept telling us the IBM had failed with OS2, it took hundreds of engineers and millions of dollars to produce a desktop OS, how could we succeed. We explained that Linux changed the paradigm. No company could obtain the economy of scale needed to compete with Microsoft and produce a desktop OS if they tried to imitate Microsoft. To compete, you needed to completely change the rules. As I had learned, a complex problem can be solved with a simple tap in the right place. No one agreed with us except Ray Noorda.
We began a skunk-works group inside Novell and actually were able to pick up some incredible talent. Over the next seven to eight months we developed a prototype of a system that introduced a layered desktop interface. The first layer was graphical. Using the browser, we developed a way to navigate the network, using pictures. If you wanted to use a traditional window desktop, you simply clicked on the image of the computer on a desk in the picture of an office that occupied the screen and a graphical windowing desktop was displayed like Windows. If you wanted to go to another department in the company, you clicked on the door that took you to an elevator and you simply selected the department you wanted to visit, like Human Relations. All the rich backend services could be exposed graphically through the browser. Once in the department, you could click on the objects you needed, like forms, and then you were able to select the appropriate form and take it back to your office. There were several objectives. We wanted to make the user experience of the network to be greatly simplified while exposing the power of the network services Novell was offering. As previously mentioned, Microsoft threatened to cut off those services because they owned the desktop. Using Linux as the underlying OS meant that we could give away the desktop if need be and at least keep Microsoft honest.
Our project brought together a number of things that would have greatly contributed to Novell’s success had they pursued them. We actually ported a fully functional web server to NetWare and had an IP/IPX gateway built in. This was all in 1993 and 1994. IPX was king of every local area network. IP was just beginning to get popular. The gateway actually provided an ideal firewall. IPX was much easier to configure and deploy than IP at the client (this was before IP was bundled with the operating system). With a web server highly optimized to run on NetWare, this gateway also acted like a natural firewall, NetWare could have gained a lot of market share as an Internet platform. NetWare was an unparalleled file server. As a Web Server, it would have been very difficult to compete with. We did not stop there however. Our team worked closely with the WordPerfect group recently bought by Novell. The concepts was to make the entire offering of WordPerfect applications “web enabled” and offer them as part of this more server centric solution. Needless to say, our project gained a lot of attention inside Novell. This was at a time when Microsoft was denying the importance of the Internet. Novell truly owned the local server market. Having released our add on projects, it could have easily transitioned that leadership into the Internet Server market.
When Ray retired and Bob Frankenburg joined Novell, Bob spent a great deal of time evaluating the work we had done. He determined, with the practical but rather poor strategic advice of the Vice Presidents that Novell should only focus on the server and not compete with Microsoft on the desktop operating system level. Furthermore, I do not think Bob recognized the immerging Internet server space. Hence, Bob did not want to continue its work with Linux or HTTP services on NetWare but focus on delivering the higher level interfaces our group had developed for the browser through the LAN Workplace team in San Jose.