Copyright 1997, Rex Ballard
When I left Loretto Heights college, I started working as a theater manager. I was working at Commonwealth theaters. It turned out that Commonwealth was restoring a theater that could be used for both films and live productions and they were hoping that I could manage that theater. Unfortunately, before the Paramount was opened, I was robbed at gunpoint. I was suspected of being involved until the robber was convicted over a year later. By that time, I was working for another theater chain.
Eventually, I moved in with my father. He had been through hell and back for nearly 3 years, and wanted to make up for the suffering he had caused in my life. I moved in with him, and started working as a salesman in a video store. At that time, I was selling VCRs but I encouraged the store owner to negotiate a contract with Warner Brothers and Paramount that would allow us to legally rent movies to customers and split the rentals with the studios. Before long, we were making more on video rentals than we were on the VCRs.
I was also getting much more interested in computers. By this time, I had taught myself BASIC, taken a programming course in college, and had worked as a data entry clerk. I had also built my own computer using 8080 and support chips and proto-boards. I was even programming them with a bank of DIP switches. It took me almost 15 minutes to load the "hexpad" entry routine. From there, I could load BASIC from paper tape. I had also spent some time at Radio Shack programming demos for the TRS-80. I eventually ended up selling computers at a video store.
At the video store, I created a number of demonstration programs, and even created some boilerplates that business owners could use to do basic data entry and record storage. In addition, we had an Alpha-Micro system and some CP/M systems. I would spend 16 hours/day in the store. I'd sell a few thousand dollars/day in computers, video games, and VCRs, and when there were no customers in the store, I would write programs.
While working at the video store, a friend fixed me up with a "blind date". She wasn't exactly "knock dead georgeous", but she was very nice, we talked well, we could be incredibly honest with each other, and she seemed to accept my sexuality. Eventually we got married. Leslie Dugan married me, her second husband, to become Leslie Ballard.
Eventually, Carl Peterson came in to see about a computer. He asked me if I would be willing to modify an accounting program he had been using to keep track of payables and disbursements. I offered to do it for $5/hour since I had no track record as a professional programmer. In a few weeks, I had made the desired changes and Carl wanted to know if I would work with some of his students. Before long, I had a small classroom and a few Atari 400 computers and an Apple ][ computer. I started writing educational programs that were consistant with the Accelerated Schools teaching philosophy (positive reienforcement, minimal attention to failure). Within a year, I had a "department" with 80 computers in 4 buildings. Each student spent 2 hours/day on a computer. In addition, each student was carefully monitored for grade level improvement. Every student jumped by at least one grade level each semester at Accelerated Schools. Eventually, the computer training became an integral part of the cirriculum and I had to be replaced by someone who had the right teaching credentials. It didn't seem right that a teacher's aide should be running the computer department for a school.
Shortly after leaving Accelerated Schools, I started working for a small start-up called Data-Law. They made an accounting and billing package for lawyers. Their first product was implemented on Series 1 using EDX. They wanted to create a similar package using CP/M computers connected together with networking software. They offered $20,000/year and I agreed to let them try me for 90 days. I very quickly took over the entire infrastructure. I started out by creating a user interface, complete with drop-down menus, listboxes, buttons, and scrolling. This was quite a feat since I was doing this on CP/M text-only systems that were written in Microsoft BASIC. Later, I developed a relational database that was more like a networked database. It didn't have all the bells and whistles of a general purpose database, but it was a good custom implementation of a relational model using referential indices.
Eventually, I got involved with the actual networking. I managed to get into the guts of CP/M, and saw some entry points that could be made public to reduce the traffic across LAN links. We turned our findings over to Novell, Digital Research, and Televideo.
Eventually, Microsoft decided to force us into an "Upgrade" that included a mandatory license fee for each user. This completely destroyed the revenue model, and since Microsoft was demanding retroactive payment as a condition of further support, Data-Law had to let me go. The were very supportive. They even let me send samples of my code to Microsoft, along with my reccomendations to CP/M. We slapped copyright notices all over everything and sent it via registered mail. The decided not to hire me, and decided to use my reccomendations in MS-DOS. The neglected to pay Data-Law for this copyrighted material. Instead, they just did a "clean room" implementation.
I ended up interviewing for a company called Computer Consoles. I was appearantly not supposed to be there, but after the interview, they got very interested. I knew a bit about UNIX, they confirmed my claims with Data-Law, I wasn't intimidated by a system diagram that featured over 200 CPUs, cluster controllers, and communications controllers. They were planning on doing directory assistance for AT&T and were going to be serving the entire U.S. The VP in charge of the project was also impressed that I would be able to move to Rochester NY in 2 weeks.
When I started working at Computer Consoles (CCI), my first assignment was to pour through the source code to the version 6 UNIX operating system, give a description of how it worked, and explain the internals. My report was limited to 30 minutes. I spent 18 hours/day for 3 weeks, and by the time I delivered the report, I had an intimate familiarity with UNIX. I gave my report, they were satisfied, and the assigned me to a series of projects that allowed me to learn their existing system and allowed me to become intimately familiar with every aspect of the system.
As a result of divestature, we had to sell the system to each of the baby bells independently. Often, we were bidding directly against IBM and at least one other Fortune 500 company. We would offer performance guarantees that appeared to be impossible. The CEO, Herm Affel, was notorious for promising what seemed to be impossible, and I became one of the key players in fulfillment. I became part of what became known as the "Special Forces" team. When a product was shipped to a customer site, we would ramp it up to production levels. Often, we had never run in that configuration, and often, we would be promising double the capacity of any system we had sold previously. When a system broke, I would be called in to fix whatever the problem was. Often, I would have direct access to source code, programming staff, system experts, and executives. I got to the point where I could use logic analysers, datascopes, sniffers, oscilloscopes, and dozens of other diagnostic tools. I also started using project management tools such as MacProject to manage and coordinate the efforts of as many as 20-30 engineers.
Ironically, I didn't see myself as a manager. I enjoyed the "hands on" experience, and I hated being the "bean counter". They never assigned me any direct reports, but instead just gave me anybody I needed in a "matrix Management" environment. In addition, I was often working on 2-3 projects at a time, making sure that they all got through the approval phase before the offers expired. Often, I would work 80 hour weeks. When I got my 5 year pin, my supervisor said it was my 10 year pin. I was getting stock options and various perks, because I would deliver.
Unfortunately, my agressive schedule was playing havoc with my family. My wife literally had to tie me down to concieve a child. Due to complications during delivery she couldn't have sex. Somehow, she was able to conceive a second child 3 years later. Finally, she decided that she wanted to move back to Colorado. As much as I loved Computer Consoles, I was getting sick of the platonic marriage, and was hoping that we could bring back the magic in Colorado.
In order to save my marriage, I took a job as a FORTH programmer for Federal Express. In terms of responsibility, it was a step down, but I still had my skills. During the interview, the handed my a SuperTracker and as I was trying it out, it reset itself. They said; "that happens all the time". I said; It looks like you need ME. They offered my the job within 20 minutes after completing the interview.
At Federal Express, I quickly distinguished my self as an amazing lunatic. I spent a day with the tracker to find out what was causing it to reset. I realised, because I was wearing leather soled shoes, that it had something to do with static discharge. I proved that this was the problem. Gene reported that the problem was static. The Director panicked and told us to fix it. We tried plastic tips, but it didn't help much. I hooked the tracker up to a logic analyser and made a small static generator. We found ways to check and correct the static RAM that enabled the tracker to continue operating after a static discharge. Essentially, it was like having a PC where you had shut off the power and turned it back on and wanted to be back to the same prompt as you had before the shutdown.
In about 3 months, we rolled out a "hard" tracker that was able to withstand not only static discharge, but also vanalism by curiurs who were late delivering a package or shipment. Eventually we went from 5% lost scans to one lost scan per MILLION. I got my first Bravo Zulu for saving the company over $1 million/day due to the reliability of the Tracker.
The REX scan and the Dimensional Weight (DIMWT) scan, both of which were assigned to me as "impossible" were completed ahead of schedule and resulted in an additional $1 million/day in new revenue as well as a similar savings resulting from the reduction of misrouted packages.
The new tracker reliability and the ability to locate misrouted packages before they were misloaded gave Federal Express the ability to reduce their lost and misrouted packages from 1% to under 5 parts per Million. The net result was that Federal Express was nominated for the Malcolm Baldridge Award. They would receive the award shortly after I left the company.
I also took on leadership of the LAN evaluation project. I coordinated several departments and several vendors to arrange staging events, demonstrations, and proof of concept trials. The system had to be capable of connecting IBM mainframes, VMS VAXen, PCs, and UNIX systems. Not only did it need to handle file and print services, it also had to handle what would eventually become known as client-server systems.
It didn't take long to figure out that TCP/IP was the best route, but I quickly learned that I not only had to select the best technology, I also had to get the alignment of directors who had very tight relationships with IBM, DEC, and even a COO who was holding a substantial interest in 3-COM. Eventually, we came up with 3 comprimise solutions including TCP/IP over Token Ring, TCP/IP over 3-COM Ethernet, and Novell with the TCP/IP extensions. The most cost-effective was TCP/IP over ethernet, but there were concerns about troubleshooting. Token-Ring was more reliable, but much more expensive. Eventually, we adopted what would twisted pair ethernet (10 Base T) which gave us the ability to localize problems, it was a bit more expensive, but we could easily extend it in the wiring closet or in the cubicle cluster.
Unfortunately, my marriage was falling apart. My wife was threatening to expose me as a cross-dresser if I didn't give her a divorce with a generous settlement. To diffuse the blackmail threat, I showed up at several holloween parties in a Maid's Costume. I made it myself since I weighed over 275 pounds at the time. We started going to a marriage councillor who basically said that even though we had the textbook definition of a healthy marriage (communication, romance, intimacy, honesty, trust,...) we were not going to be able to have mutually satisfying sex with each other. I chose to be celebate, Leslie chose to have an affair, and her boyfriend decided he wanted it ALL. Leslie waited until after I celebrated my 9th year of sobriety and announced that she was engaged. I forgave her for the affair, told her she would have to wait one year and that I would forgive her if she changed her mind.
It turned out that the VP had been getting reports of my activities and had decided that I should not be at Federal Express because I was a "pervert" because of my Halloween costume. To make matters worse, my boss showed up at an N.A. meeting. I arrived late, but I saw that he had received a newcomer's chip. I had become a problem for too many people.
I kept hoping that this would all blow over. What I didn't know was that anything I shared at an A.A. or N.A. meeting was being relayed directly back to the VP of engineering. Todd would tell his brother who would tell my coworker who would chat while riding the plane with the VP. Since she was from Nashville, she would get a seat on the jet whenever it was available.
Eventually, after nearly a year of "councilling sessions" with my immediate supervisor and what could easily have been proven to be harassment. I was told that if I doubled the revenue of Federal Express in the next 90 days, I would still be asked to leave. I left the company quietly. Leslie and the kids had already moved in with Jerry. I moved up to Denver.
I joined Great West Life about a month later. It was a great company. I was responsible for intergating the business units, including the technology. The wanted to migrate some of the applications from 3270 mainframe terminals to PC and UNIX platforms. I created a number of gateways, including an FTP client that would allow mainframe files to be fed into the standard input of a UNIX process. It became very easy to concurrently run dozens of analysis programs against a single pull of the files. We then started writing RPC calls against VSAM and DB/2 databases. Soon we were feeding transactions into IMS - it was, in essense a "message Queue" that provided near-real-time response. Before long, there was a "bidding war" and I ended up working for Jim Rosborough in the 401K division.
Jim was a great leader. He was chairman of the LOMA infrastructure and architecture committee. He was very interested in dealing with some of the problems associated with large-scale distributed processing. I soon found myself leading a team of 7 programmers. We set up network management, monitoring, and warning systems. We set up multitiered distributed processing systems that allowed us to farm complex operations across multiple systems and included the checkpoint and recovery mechanisms to assure that the processing was completed correctly. I also led the effort to develop an interactive voice response system used to allow 401K customers to select their mutual funds by phone. It was one of the first IVR systems in that Industry.
Jim also had me evaluate some SMP systems, including Sequent, Pyramid, and Sun systems. The Pyramid system was very powerful. They had the source code to Oracle and were able to reduce our average response time from 5 minutes to 45 seconds. We were afraid the $1/4 million price tag might be too high, but the VP had just been given an estimate of $30 million to upgrade the mainframes. We were able to offload enough of the system to Pyramid and Sun systems that we were able to save the company nearly $25 million in less than 3 months.