Most careers, I believe, have setbacks. And many people go through jobs that at some point seem to require compromises they don't like. In my case these problems were relatively brief, and, in any case, highly educational.
In 1969, I got my first "real" job, assistant professor at a state university. My PhD was in pure mathematics, and the post-Sputnik push of government in the sciences meant there was still growing enrollment in mathematics; the department was hiring several young people a year.
But when I was considered for tenure in 1974, the situation was very different. Enrollment was shrinking, the department faculty was shrinking, and several young faculty, including me, were denied tenure and would have to find other jobs.
The job market was starting to look better in computer science than in mathematics. I'd had some summer jobs working with computers. Was it possible to make the change? I asked the chairman of our newly formed computer science department if one could become a computer science professor without ever having taken a computer science class. His answer was short: "I did."
Computer classes were rare when he and I were students.
As I started job hunting, the computer science chairman had a call from the university's College of Business: Could he suggest someone for the job of computer jockey in the College of Business? I took the job, at some cut in pay and status.
My main task there was unexpected. The College of Business had a contract from the state government to provide services for the State Council of Economic Advisors. We made revenue forecasts for the state budget process, economic tables to help recruit industry to the state, and tables for state agency annual reports.
Computer programming was needed, but there was a much broader opportunity to use my mathematical and statistical knowledge, as well as any stray side interests I could bring to bear.
The notion of a "database management system" was brand new to the state government, and asking people to think in those terms was a hard job even for an experienced teacher. I spent a month, early on, walking around state agency offices asking "What numbers do you produce here? What numbers do you get from elsewhere?" and "What numbers do you wish you had, but can't get?" I then tried to organize as many of those numbers as I could in some rational fashion.
The numbers that governments use often have much more complex definitions than one might expect, however.
I once asked the appropriate official in Washington, D.C., what the unemployment rate was that week.
"Just under 6 percent, nationally," he said, "but don't ask which states have a rate under 6 percent, because there aren't any." He said there was political benefit in keeping the national rate low, but the states received more federal aid if they had a higher rate. So the definitions were different at the federal and state levels. Members of the United States military counted as "employed," nationally, but not within any state. Persons unemployed more than six months might be considered "unemployable" nationally (and hence no longer counted as unemployed), but might be counted as unemployed in a state if they were still drawing unemployment pay.
This was not a case of twisting the statistics to come out the way one wanted. Firm rules had to be followed, it was just that those rules had evolved by taking into account differing goals and purposes. It took time for me to understand those systems, and even more time to figure out how to work with them.
I became very popular with some local officials, because I was very good at figuring out which numbers made their towns look good - or, sometimes, bad. They could get extra federal aid if their town was in the poorest 10 percent, something that could be established in a variety of ways: If they were not among the poorest towns on the Department of Labor list, they might be low on the Department of Commerce list, or high in what percentage of income was from welfare payments.
But I was very unpopular with other officials, because the professor in me usually wouldn't sign off on a number until I'd made an honest effort to understand what it meant. Someone had to say to me, "I need the number Tuesday, and I don't care if it is right. I just care if it is Tuesday," as well as "I asked what time it is, I didn't ask how the watch works."
I worked on an important state report for months. I expected my name to appear as its author. But I felt too much pressure from my superiors: They wanted the numbers come out a certain way. Finally, I issued the report - with no author listed. To a budding professor, giving up credit for a publication seemed a high price.
After two years, I was ready to find another job. I returned to teaching and won tenure as a computer science teacher in a comparable university. For many years in my classes afterward, I used the experience and examples I'd gathered in my two years working with state government. My students have come back after graduation to say how valuable those examples were.
In hindsight, the university that denied me tenure did me, and my later students, a considerable favor.