|The picture above
is from about
1957-8 winter. Snow topped Organ Mountains, A Mountain,
and many other buildings visible. Oddly, I was always
impressed by the
foreground plowed field detail (not really visible in this
only in the old Kodachrome slide).
The scene at left is a familiar sight to anyone who has attended New Mexico State University in Las Cruces. It is Goddard Hall, which was the engineering building when I began my education there in 1956. I ran my first computer program in the Goddard Hall annex in 1958 on the Bell Labs Model V relay computer, thanks to Heintz Gehlhaar, an electrical engineering student, and Professor Harold A. Brown, head of the EE department. Heintz put the thing together after it had been delivered there from Ft. Bliss where it had been a ballistics computer. Professor Brown allowed me to use it. [I previously had referred to this as a Model 4, but recent postings of history reminded me it was a Model V.] This began a long career in computer software development during which I worked with many interesting and wonderful people.
The biggest compliment of my professional life was paid to me by Howard Aiken, the developer of one of the first digital computers (see "Mathematics and Computers", by Stibitz and Larrivee, McGraw-Hill, 1957, pg 53) at Harvard University where the computer center is now named after him. Grace Hopper worked for him at the beginning of her illustrious career in 1944. In 1969 at a party given by Maury Halstead, who had invited Howard to Purdue for a colloquium, Howard approached me and said "Bill, Maury says that you are a good systems man. If you ever need a job, let me know". Unfortunately, Howard died a few years later in 1973 (and before I needed a job).
Maury, himself, was a true pioneer in several areas. After being labeled the "grand kahuna" in Hawaii, because of the successful cloud seeding experiments he did there in the '40s, he moved to the U.S. Navy Electronics Laboratory (NEL) in San Diego after a period at Texas A&M where he was a professor of meteorology. With Harry Huskey of the U. of California in Berkeley (president of the Association for Computing Machinery (ACM) in 1960-61), the NELIAC language was developed at NEL. NELIAC was the first self-compiling compiler and one of the first implementations of a higher level language (See: "Machine-Independent Computer Programming", Maurice H.Halstead, Spartan Press, 1962). Later, Maury became interested in the mental effort required to develop computer programs. This resulted in the work he first called "Software Thermodynamics" and later labeled "Software Science". Those who know this work probably know it as "Halstead Metrics"
I had several jobs here. One was as a "human computer" executing programs on a Friden mechanical desk calculator. Another job was with a "real" computer, the Bell Labs Model V, which I programmed sitting at a teletype that punched paper tape. The paper tape was later fed to a computer that was crafted from 10,000 telephone relays. With 15 storage registers as its total memory, it had less capability than any of today's programmable hand calculators. It would multiply two floating point numbers (yes, floating point) in 2-3 seconds (seconds!) and a square root in about 6-7. But, it was impressive, taking up two very large rooms, and having its own DC generator and power control panel with warning bells, big throw switch, and the whole works.
This was my first job after college. I had graduated with an M.S. in mathematics, and that helped me get the job. I had worked there the summer before and had reported to Dr. Maurice H. Halstead, with whom I thereafter had a long association in the computer field. I worked on an analog compiler and various peripheral processor utilities. I was there just under a year. Maury went to Lockheed in the fall of '63 and I followed him there in April '64.
At Lockheed, I was in Maury Halstead's advanced software development group. We worked on the IBM 7094 Neliac compiler, moved it to a Univac 1107 (then to 1108). Also worked on the IBM 7094 to Univac 1108 Decompiler to Neliac. There were many interesting projects in this period including the first copyrighted programs (the Lockheed Blood Bank program), converting the weather forecasting programs of the U.S. Air Force Global Weather Forecasting system at Offutt AFB., work on the forerunner of the Dialog Computer Services information retrieval system (which was later spun off). After a short time as the supervisor of the Automatic Test Equipment software group for the Poseidon missile system, I finally had enough of the defense industry and began to look elsewhere.
This endeavor began as Teknatronics, Inc. and later became Combinatorics, Inc. Maury Halstead had gone to Purdue as a professor in computer science. I had called him to get a reference in my job hunt from Lockheed. He invited me to join a startup in Lafayette, IN. So, off we went to Indiana, a place we had never visited (I didn't even go there on an interview trip). I won't bore you with the details of this venture, but I was the chief cook and bottle washer (and ultimately, vice-president). Finally, Maury Halstead and I landed a contract with the Univac Division of Sperry Rand in Roseville, MN to help design and develop an inverse compiler for one of their aging computer systems (Univac 494). That is how I wound up in Minnesota.
Beginning with a continuation of the inverse compiler project, I worked on software engineering tools (including an adapted version of ISDOS project software), rapid prototyping, and finally, personal computer software and the integration of PC software with mainframes.
Sperry opened a new development center in Atlanta, GA in mid-1985. I joined the Corporate Technology Center (CTC) there in the fall of 1985. The CTC was chartered with the system integration of a showcase of hardware and related software engineering tools for the corporation and for customers. A second group was chartered to develop a secure operating system for Sperry mainframes and a new transaction system for the new OS. In mid-1986, Burroughs merged with Sperry forming Unisys. The center was closed being deemed as unnecessary in the merged company. One of the best groups of software developers from all over the U.S. was disbanded. Some formed new companies, others went back to their previous companies or to different companies.
From Atlanta, I went to Gould as the project scientist for the secure Unix project there. This group had, prior to my arrival, completed the evaluation of the first "secure" Unix system at the C2 level. After learning a lot about Unix, and some about security, I moved back to Unisys in Roseville. The group in Urbana was purchased shortly thereafter by Motorola as Gould was sold off in parts. Many of the folks that I worked with there are still at the Urbana site.
Back in Minnesota after two and a half years, and three home sales and purchases later, I began in software development again, first on Xenix, and then on Unix with a smattering of mainframe. In the last few years, I represented Unisys Roseville on the X/Open Distributed Transaction Processing Working Group. This was very interesting, if not very productive. Meetings were held every 2 months alternating in Europe, U.S., and Japan. At the same time, I consulted on matters related to that work with Unisys development groups. As my last assignment before retiring, I was involved in porting a transaction system from Unix to Microsoft Windows NT. That did me in, I guess, because I decided to retire after completing my initial assignment that included the design of the protected subsystem, and the implementation of Unix style message queues for Windows NT.