Dan can be contacted at [email protected]
I remember the reasons that I moved from Pennsylvania to Miami with great clarity. My wife and I were going from the gray and brown to the green and bright and the trip to Florida in 1971 has become family legend. Two years later, our expanded family (our daughter was born in July 1972) got back in the car and moved to Boston. I don't remember exactly why. I know that I had taken a job as the Director of Data Processing for Massachusetts General Life Insurance Company, a small insurer that occupied a few floors of an office building on Federal Street, but the rest is a fog. I remember looking out of a window onto Federal Street on a blustery and overcast day in May when a co-worker explained that there was only one good day of weather in Boston each year. Unfortunately, he remarked, we had already had it. I remember the reasons that I left Boston to return to Jacksonville Florida in 1975 but I'm just not sure why I went north in the first place.
While I was driving up and down the eastern seaboard something happened to the data processing business. It changed.
In 1974, IBM announced virtual storage and the MVS version of OS/360. With virtual storage, a programmer could request an address space larger than the physical memory installed on the computer. At Mass General Life, we didn't have our own computer but we rented time on the machine up Federal Street at the Boston Stock Exchange. Accordingly, we were the last on the list of companies that the IBM salesman visited to discuss the announcement.
"Why should a programmer use more memory than the mainframe supports?" I asked him.
"Why, indeed!" he responded with the usual IBM enthusiasm. "Just think of all the new applications that you can write. Imagine how many partitions can be active at the same time."
"How does it work?" I wondered.
"Well, if there isn't enough memory, the operating system will write a portion of the program or a segment of data to disk. Then, when that portion is needed, it will read it back into storage," the IBM salesman lectured.
I have to admit that the concept of virtual storage, later called virtual memory, was lost on me. We were locked into the sequential processing of large master files. The system couldn't validate transactions until they were sorted. It couldn't change policies until the transactions were validated. It wouldn't apply premium payments until the policies were changed. One step after another, each a part of the daily processing cycle. Why have multiple partitions when you could only do one thing at a time, anyway? And what program couldn't be written in 64K or less?
At about the same time, I got a note from our company's president to return a call from a salesman for Data General, a high-tech company located out on Route 128. Although the salesman wanted me to visit their facility, we agreed on a lunch meeting the following week. Data General was started by a few engineers from DEC who believed that they could build a mini-computer that would appeal to businesses based on price and performance. They first built a machine called the Nova and the salesman wanted to tell me about their new offering, the Super Nova.
The Data General salesman was a true believer and I was a confirmed skeptic.
"Have you heard that our new COBOL compiler is faster and more efficient than IBM's compiler?" he inquired.
"Who's going to rewrite all of our programs from assembler language to take advantage of your new COBOL compiler?" I asked.
"You can lease a Super Nova with tapes and disk for less than your rent at the stock exchange. Just take those savings and hire a few programmers for the rewrite." To the Data General salesman, it was a straightforward decision.
I remained unconvinced and recommended against the switch from a mainframe to a mini-computer. We were not looking for something smaller. At the time, our holy grail was a meg and a mike. If we could get a machine with a million positions of storage (a meg) and a microsecond processor speed (a mike), we could do just about anything. After all, everyone knew that bigger was better.
Data General was a good company with good people. They had a vision of data processing, however, that would take a quarter-century to become reality. During the '70s, the momentum belonged to the mainframe and it would take the recession of the '80s and the success of the personal computer to make people change direction. Eventually, Data General began to manufacture storage systems before they were purchased by EMC in 1999.
As my family and I traveled from Boston to Jacksonville in the late summer of 1975, I was completing my first 10 years as a data processor. It was a decade when computer programming had come of age. Although the term was not then in general use, we programmers had become the masters of the universe. Since standards were non-existent, we used whatever languages and techniques we liked best. Since users were intimidated by us, we dictated the requirements of the programs that we wrote. We made our own rules and we programmed on our own terms.
Unlike other departments within the company, the data processing department had an existence beyond corporate boundaries. Other than the operating system and a few utilities, software was custom-written or, at least, customized to a unique specification. When a programmer was assigned to write a program in the company's primary processing system, it was a job for life. In 1975, we performed our own analysis, wrote our own programs from beginning to end, and performed our own testing.
Although these programming tasks were solitary, there was community. A gathering of programmers was like a meeting of sovereigns. Each data processing shop was a separate principality and its director was the lord of the manor. We were a heterogeneous group of white males who saw our profession as the equal to medicine or the law. We would confer but we would not become subject. In all of this, our behavior was not malicious, just uninformed. It was our arrogance as much as the technology changes that would shortly come that sowed the seeds of our undoing. The coming changes, however, were only obvious in retrospect.
In November 1975, another young man left Boston and returned to his home in Seattle. Before the year was over, Bill Gates, having dropped out of Harvard, formed a small company with his partner Paul Allen, to follow his dream of placing a computer on every desktop. They named their company "Micro-Soft."
- A Personal History of Systems and Computers: Part 1
- A Personal History of Systems and Computers: Part 2
- A Personal History of Systems and Computers: Part 3
- A Personal History of Systems and Computers: Part 4
- A Personal History of Systems and Computers: Part 5
- A Personal History of Systems and Computers: Part 6
- A Personal History of Systems and Computers: Part 7
- A Personal History of Systems and Computers: Part 8
- A Personal History of Systems and Computers: Part 9