like I need another hole in my head. But I’m plunging in anyway.
Hello there. My name is David Bryant. Currently I’m living in Canyon Lake, Texas, with my wife Kathryn, my pug Tudie, my papillon Justin, and two parrots named Spanky (she’s a little rascal) and Boom-Boom (don’t ask why). Here are some pictures. Sorry, I don’t have any pictures of the birds handy. Maybe another day.
And here’s a picture of Tudie the pug. I’m curious — how can I get these to display alongside each other? Research into WordPress internals is necessary, I guess.
I don’t have time to tell my whole life story, and you’re probably not interested anyway. So I’ll sketch this out pretty swiftly. I was born in Palmer, Alaska in 1951. I attended all 12 years of public school in Palmer, and graduated (first in my class) in May, 1969. I attended the California Institute of Technology in Pasadena, California, and received my Bachelor’s degree (BS(Ma)) in June, 1973. Since I was a math major and I was not interested in an academic career, I entered the actuarial profession, obtaining the professional designation Associate, Society of Actuaries (ASA) in November of 1974.
I worked for several insurance companies in Los Angeles while I was earning the ASA designation. I soon tired of the crowded freeways and the perpetual summer in southern California; in September of 1975 I started working for United American Life in Denver, Colorado. I liked the job at UAL, but good things never last long enough. UAL was acquired by Western Preferred Life in September of 1977. By March of 1978 I had had it with WPL’s dishonest — crooked as a dog’s hind leg is what I ought to say — management team and their fraudulent accounting practices. I resigned my position and took my concerns to the Colorado Insurance Department’s actuary, Anthony Fagiano. An investigation ensued, and in 1984 Western Preferred was put into conservatorship; a couple of the executive officers even went to jail.
In September, 1978 I went to work for Western Farm Bureau Life Insurance Company (WFBLIC) as a computer programmer. I had already developed strong programming skills with IBM assembly language. Since WFBLIC was installing LIFE-COMM from Equimatics in Dallas, it was a good fit. While working at WFBLIC I fell in love with Kathryn, who was also an assembly language programmer. She left WFBLIC for Security Life of Denver in April of 1984. In July we got married, and in late August we bought the house at 520 South Corona Street, where we lived happily — if not always harmoniously — for almost 29 years.
While working for WFBLIC I eventually lost the sour taste that actuarial work at Western Preferred had left in my mouth. In mid 1983 I joined WFBLIC’s actuarial department. That got stale; in January of 1985 I went to work for ICH in the Denver Tech Center. That didn’t last long. ICH was almost as crooked as Western Preferred had been. Rather than hang around and get burned again, I joined the actuarial department at Security Life of Denver in March, 1985. At last I had found a really good actuarial job! Security Life was an A+ company with a distinguished history going back to 1935. And Dick Horn, the president, was an actuary with an excellent professional reputation.
My assignment at Security Life allowed me to combine my actuarial and programming skills: I was supposed to write a computer program to project the future cash flows and profits / losses expected to arise from the company’s existing block of business. At the time I started working there, Security Life was still running OS/MVT on a 360/50. After running a bunch of benchmark tests using both floating point and fixed point binary calculation routines, I decided to go with fixed point binary; the same calculations took 2½ times as long to execute when coded in floating point.
It took about a year of intense effort to build the new projection program. The old program had been written in COBOL and took 20 – 30 hours to project 30 years of future cash flows, profits, and losses. My new program got the job done in 5 – 6 minutes, a 99.5% reduction in run time. Of course, the fact that the old 360/50 had been replaced with a shiny new 4381 didn’t hurt the run time. But the main improvement I had made was to eliminate most of the I/O operations the old program performed. Instead of reading lots of rate files, I calculated cash values and reserves for individual policies on the fly. Disk I/O operations are about 1,000 to 10,000 times slower than the internal processing speed of a modern CPU; you can increase the CPU time quite a lot (say 400% to 500%) and, if you effect even a small reduction in total I/O operations, the program will run in much less elapsed time.