The Start Ups
Information Nexus, Ltd.By 1980 there was beginning to be some interest in using Unix for something other than plain ole research. So, Matt left the easy University life and started a business (!). The company was staffed by Matt and his wife, Ana, and another technical fellow and his wife. Later, it got to include two honest-to-goodness employees. That was the best move he ever made. It turned out to be a very broadening experience and made him a lot of contacts in the fledgling industry (Unix/Workstations) throughout the US and even in Asia.
The business went through various transformations. They build shrink-wrap products; hired out as consultants; built customized systems; they even resold hardware (Onyx, Zilog, Codata).
By a series of compounded mistakes, that little company managed to get up to nearly $1M in revenue (!).
Henry Ford Hospital(s)The first thing Matt did with his new business was to work on a project for Henry Ford Hospital in Detroit. Their radiology department wanted to put computers in the department to better keep track of their case folders (i.e. where they kept all the radiology records for a patient). Henry Ford itself is a pretty big hospital with over 900 beds and lots of out-patients going through it; but in addition, they had seven clinics spread throughout the greater Detroit area and its surroundings. They kept seven warehouses full of Radiology case folders. It was a miracle that any folder could be located at any one time. They knew they needed a better system to keep track of these things, and they figured that they needed computers to do it with. Instead, Matt and his associates decided to first do a Structured Analysis of their operations using Data Flow Diagrams, to see how folders flowed through the system. Structured Analysis using Data Flow Diagrams was at that time a new technique created by Tom DeMarco and published and promoted by Yourdon, Inc.
In the end, they were of the opinion that the fundamental problem was their process which had developed over the years, as the hospital grew, and it was never designed. The solution had more to do with redesigning their processes and procedure than adding computers to the mess. The customer didn't like that answer and so Matt et al immediately specified a computer system configuration to take care of the problem. Thus began Matt's education in the real world of business.
Probably the most memorable and rewarding experience Matt had during this period was his involvement with a customer in Hong Kong.
One thing led to another, and by 1983 he, too, rolled west. He moved to San José, California, the equivalent of Mecca for nerds-cum-entrepreneurs. He went through two start ups before joining Sun Microsystems in 1986. One of the start ups (Codata) made Sun workstations before Sun was even founded! (i.e. the system was based on the original CPU design that Andy Bechtolsheim did for his thesis for the Stanford University Network project.)
Sabers claim to fame, so to speak, was their very high resolution, color display. They built it from the ground up. Well, almost. The company bought the glass tubes from a japanese manufacturer. But that was it. The rest of the display was designed and built by Saber.
Saber°s display technology was way ahead of its time. Even today, current display technology is not quite up to the resolution of the Saber display. That technology was capable of displaying 1640 by 1280 pixels, with 24-bits of color. That°s over 2M pixels, and more than 6M bytes per frame! The system operated at 72 frames per second, or over 432M bytes per second. Actually moving those many bytes into the display was not the hardest problem (although it wasn't trivial, either). Switching the color guns on and off was the realy hassle. Saber solved it by designing a hybrid circuit that was able to do the job. Nanao manufactured the hybrid for Saber.
Matt started as Director of Software. But soon after became Director of Engineering, responsible for all aspects of product development, including the system software and hardware, as well as development of the display subsystem itself.