I wrote my first computer program in 1967, as a college student. Back then, I think there might have been one or two CRTs on campus. They were reserved for computer science grad students. The rest of us had to punch Hollerith cards. A simple Cobol program would take nearly an entire box of cards. If I remember correctly, those boxes were about 24 inches long. Heaven forbid dropping the box before the program was debugged and the cards were sequenced. Now that's old technology.

For the next many years, all of my serious programming was done on large mainframe computers - Control Data Corporation Cybers for those who are old enough to remember those powerful old machines. They were used in the early days of the Manhattan Project, and NASA's space program. They are extinct now. I'm getting there. I actually went to work for CDC as a tech support person. My main job was to keep sales people from making promises they couldn't keep. I was so good at the technology they made me a sales person. That was when the smallest Cyber sold for over a million dollars and wasn't as powerful as my current smart phone. When PCs started gaining popularity with engineers and scientists, it was common to hear CDC people say: "You can't trust a computer you can see over the top of." CDC wen out of business several years ago, and the main applications that were popular are now vastly improved, and running on personal computers.

When PCs came out, I wrote a few programs in DOS, and played with a Comodore 64. However, my first serious PC programming was to develop an office automation program for my small business. It was just for internal use until others offered to buy it for their offices.

Once I started selling the product, I was obligated to release enhancements. Users made suggestions that resulted in many new features and user interface changes. It evolved into a single product with multiple interrelated applications. The combined size is over 3.5 million lines of Delphi 2007 code.

As you can guess, I'm stuck with D2007. It just doesn't make economic sense to try and rewrite all that code into a later version. The end result would look and behave fundamentally the same as it does right now. That's both good and bad. The good is, I don't have to do all that work. The bad is, I don't get to advance my skills by skipping all that work.