In 1970, the year of my birth, my father was a Merrill Lynch stockbroker by day and a waterfront security guard by night. In the mid-1980s, he saw that I had some talent with computers. Maybe I could learn the markets and make it big on Wall Street. It also would have been fine if I had turned out to be the next Bill Gates.

Dad asked me for a stock pick. His favorite strategy was to buy out-of-the-money call options that would pay off if the stock went up. I found a story in the Business section of the New York Times about Western Union (NYSE: WU). Since Western Union had been sending money over the wires for so long, I figured it would lead the way in sending money over computer networks. Instead, the company went bankrupt. Dad never asked me to pick stocks again.

I learned how to program computers, but my main activity was hanging out on BBSes and IRL with other computer users in my local calling area. When I was 16, I met up with Doc, a graduate of a Caribbean medical school who got into the database business while trying to pass his licensure exam. He was an excellent salesman, a competent programmer, and a good guy, very generous with his time. I spent an entire summer working in the basement of an actual video store trying to write software to run video stores. That was beyond my capabilities, but I was very good at technical documentation and picking apart other people’s code. I liked reading the manuals.

After college, I was living at home recovering from my Carnegie Mellon experience. I went back to working for Doc, but not for long. My father wanted me to get a regular job. I found one through the Classified section of the New York Times. A company was looking for someone who knew the quirky Macintosh database software that I had learned working with Doc. I got the job, and it paid well with good benefits.

But Doc wasn’t happy, saying I had poached his prospect. He had gone after that company’s business, but they weren’t interested in hiring a consultant and never did business with him. I had no idea about any prior contact when I answered the classified ad. Doc might have caused me some legal headaches out of spite based on some papers I may have signed while heavily medicated, but like I said, Doc was a good guy. He let it go without lawyering up. But we were done.

Competitrack was founded in 1987 just before Black Monday. Bob was working for Chemical Bank gathering data on which competitors’ ads were being placed in which New York newspapers and magazines. After doing this thankless task for a while, he realized that he had the seed for a startup business. He hired a staff to track advertising in print and on television, wrote a database to capture the details, and then generated customized reports for ad agencies representing every competitor in an industry. Bob built the company on one of the few relational database management systems you could find on the Mac at that time, and in 1993, he hired two programmers – me and Dave – to work on the code.

We spent about a year learning and optimizing the legacy database, which had the fatal flaw of being built using a proprietary data format with no APIs or other external hooks. If you wanted to work with the data, you had to write code in the proprietary database application. Or, you could export the data into a separate file, and even on the fastest Macs, the import-export process was very slow.

The second year, I worked mainly on ancillary solutions not directly tied to the legacy code. I built an engine for estimating the cost of advertising campaigns. I figured out, without the benefit of documentation or search engines, how to read Nielsen data from EBCDIC-format mainframe reel-to-reel tapes. And I designed and rolled out parallel Mac- and PC-based video transmission services allowing us to send digital recordings of television advertisements to ad agencies, and this was well before video sharing was commonplace, around the time of the Netscape IPO.

The third year, the company brought in someone more experienced to run the IT department. Even while keeping the legacy software updated with new feature requests, we tried to rebuild the legacy system using an Oracle back-end database. This would free the company forever from our dead-end legacy software, but it was the classic dilemma of having to rebuild an engine while still flying the plane.

I built a new system, but management didn’t want to bet the company that I could bring the alpha version to a production-ready state. The project was put on ice indefinitely, or at least until they could bring on more resources. This was a major disappointment. I was hoping to upgrade my skills by getting out of the legacy trap and had believed Oracle to be the bridge for whatever I wanted to do next, whether it was another job or starting my own company.

Then, I had to give up my Bleecker Street apartment, which was being sold. Feeling uncertain about my prospects at the company and beyond that, my future as a programmer, I moved back to New Jersey to live with my parents and plan my next move.