Graham Seibert Autobiography draft Jan 15, 2013 Page
Download 412.6 Kb.
|
Independent ConsultingI quit Booz to become an independent consultant shortly after returning from Argentina. It was time to become fully self-sufficient. I had no consulting prospects, but I had a pretty good collection of skills, a real estate portfolio that was at least paying for itself, and a decent stock portfolio. It was time to take more risk. At the time I left Booz I had no prospective clients and no useful contacts. I started making cold calls. When I phoned my friends still at IBM to let them know what I was up to, it turns out that they knew a few people who were doing what I was attempting, and they put me in touch. By 1981 I was a founding officer of the Washington Independent Computer Consultants Association, WICCA. Not to be confused with pagan witchcraft. The association with other independent consultants was valuable because all of us had some useful contacts, few of us did exactly the same thing, and we could team up on projects that required several people. Also, sooner rather than later, a group of brokers emerged whose speciality was matching consultants with opportunities. I worked a number of small jobs, gaining experience and building a reputation. After our son Jack was born in April of 1982 Mary Ann stayed home. It was in the early days of the IBM PC. Hard drives had just become available. I designed a Military Inventory Control System to do the same job that Major Glidden and I had automated seven years earlier in Frankfurt. I naïvely thought that because the military needed it, and they could simply license it as needed, they would buy it. I asked Mary Ann to help program it. Our office was in the basement of 2114 Huidekoper. We kept it there even after we moved to 2120 Huidekoper. We rented out the top two floors and kept the office and garage. Mary Ann was not exactly a willing worker. She didn't like to be in the basement and she didn't like the undefined nature of the project – writing software on spec. She wanted precise instruction. We eventually sold a couple of licenses to the Army in Korea and the Navy in Puerto Rico. It involved both of us going to Korea and me going to Roosevelt Roads, PR. By that time I was doing quite well as an hourly-rate consultant, and had a much better appreciation of how political the procurement process is. Given Mary Ann's lack of support, the daunting obstacles selling to the military, and my succcess in pure consulting, we dropped the project. Mary Ann went back to work for Booz, Allen about the time Naomi entered nursery school, fall of 1986. Until that point she and the babysitters, Tan and Lourdes, had pretty much taken care of the kids. I played with them and read to them, but I wasn't needed and wasn't very often asked to feed, bathe, change and amuse them. I had lots of other things to do: build a business, write my first book, act as a leader in church and professional organizations, and look after investments and real estate, among other things. I could not foresee where this would lead, but leaving discipline and character formation up to the women appears in hindsight to have been a mistake. Computer Sciences Corporation (CSC) was my first major consulting success, about 1984. The introduction was casual – they needed somebody who could tell them about the 3741 programmable workstation, the machine I had programmed for the Army, and write a few words about IBM’s midrange computers for an upcoming proposal to the Saudi Ministry of Petroleum and Minerals. As it turned out, however, the head writer for the proposal was stricken with a medical problem. I was drafted to lead that effort, and a couple more. Altogether PETROMIN released six requests for proposal (RFPs) and CSC bid on four of them, winning three. After the victory parties for the three proposals that Computer Sciences had won, things went quiet for a while. Between the award of the contracts and the start of the work there was a several month break. Computer Sciences assembled teams to handle each of the projects. There were essentially three of them. One of them, called Tools And Methodologies ,set out methodologies, operating procedures and so on for the Saudi Ministry of Petroleum and Minerals (Petromin). The second set up a machine and operating system environment. They needed to install the IBM 370 computers that would do the job in Jeddah and Riyadh. The third project involves installing accounting software. The Computer Sciences team had taken a conservative approach, purchasing a license to the McCormack and Dodge Millennium suite of software. It was standard accounting software written in the COBOL language. They would simply implement it in Arabic, using the tools that had been procured under the tools and systems contracts. They put a young PhD in charge of the project. He ignored the approach we had written into the proposal. I had proposed a sandwich-like approach. We would keep the McCormack and Dodge programs as close to the original as we could, using an automated approach to modify them on the front end to accept Arabic input and the back end to produce Arabic reports. The entire philosophy was that we would not touch or attempt to understand the logic of the programs. McCormack and Dodge had invested hundreds of man years in developing the logic of those programs and it was important in my mind that we leave them absolutely alone. Their detail-level documentation was scant, and reverse engineering programs to determine their logic is a very difficult, tedious task. We did not have a need, or the resources, to attempt it. The project team ignored the proposal, and opened up the COBOL programs and had regular ordinary programmers attempt to convert them to operate in Arabic. The effort was not a success. They were several months behind schedule and about $10 million in the hole when in desperation the management team of John Harned and Reg Boudinot gave me a call. We had a council of war in which I held firm to the position that we should do it do it over, and do it the right way from the beginning. If we changed the logic of the programs in any way, the vendor would not be able to support us or Petromin, and we would be at a loss attempting to implement version changes coming from M&D. This approach required two programs. First, on the front end, it required programs that would look at the terminal input, the the cathode-ray tube display input to the system and convert those to operate in Arabic using the special Arabic language terminals which Computer Sciences had procured under the other project. The screen support was written in a separate language, CICS, which delivered data to the COBOL programs, one record at a time, in the same format as if they had been read from a tape drive or a disk. The second program, the one I wrote, combed through the programs looking for everyplace that produced Arabic output for the printer. It had to identify every literal (bit of fixed text) which needed to be translated into Arabic, and to ensure that even text that got moved around a few times before printing was appropriately translated. The translation itself was to be done by language experts, not programmers, using the lists of literals we pulled out of the English-language programs. We needed two programmers, a CICS expert to convert the screens, and one to convert reports. I knew a genius programmer named George Halberg and I brought him in the project for the former, and I did the reports myself. There was considerable anxiety, and a lot of pressure, as George and I took something over a month to program our stuff. However, once we finished, they were able to convert the programs by our “sausage maker” process. We simply put the original McCormack and Dodge program in one in end and got a converted program out the other, including a list of literals that needed to be translated, and fields that would be carried in both languages. It was the automated process I had written about in the proposal. This approach worked, and we became local heroes. Expensive local heroes, however, and as always happens in the business of consulting, the measure of our success was that after a few months we were once again out of work. I did a couple of small jobs, continuing to work through the established brokers, people who found work for independent consultants. Armand Posner referred me to Neil Grundstra for a job with Honeywell Federal systems. They were implementing a cost modeling system similar to the one that I had written for IBM in 1979. I came up with a proposal for how to do it. They wanted to implement in software within the Honeywell suite. This involved the public domain UNIX operating system and the C language, using the Oracle database. I did not have experience with any of them. In a leap of faith on their part, and a stretch on my part, I wound up leading the project. It was successful, and along the way I learned those three pieces of software quite well. It was a real eye-opener to me. I learned why IBM had faded so badly in the marketplace. IBM’s solutions, their languages, databases, and operating systems, were simply not as good as the competition. In this case, the freeware competition. UNIX, developed in a university environment (Berkeley) and under the auspices of AT&T, was a groupware project. One of the first, and still one of the most successful. Download 412.6 Kb. Share with your friends:
|