Connecting with old friends on Facebook lately led me to remember those early days as a teenager writing video games. It's incredible how much has changed since then. It was nearly impossible to get any information on how to program a computer, much less on how to write games. The library books were typically user and application-oriented. C and assembly books could only be had at the bookstore at ~$60 a piece. In the days before bookstores came with sitting/reading furniture, being a kid without a lot of cash, I'd go to Waldenbooks in the mall and stand there reading, absorbing as much as possible before I had to leave. Today, on the other hand, there's rarely a need to buy technical books because whatever I need can be found on the web faster and for free. Today is definitely better.
With a tiny bit of programming in school (Applesoft and Commodore BASIC) when we got our first computer (a Tandy 1000SX), I naturally started with GW-BASIC. It was too slow for the kinds of games I wanted to make - games like Ultima V and Times of Lore.
For Christmas, I asked my parents for a BASIC Compiler (a what?, they asked). It was what I'd circled for them in a magazine - Turbo Basic. I expected to drop my games in and have them run like crazy. There was a hitch, however. The Tandy 1000SX shipped with an OEM version of Microsoft's GW-BASIC that natively supported the Tandy's enhanced CGA video adapter (essentially the same as the the IBM PCjr's with the 16-color 320x200 mode). Turbo Basic didn't. So, as long as the games could be confined to the horrible four-color plain CGA, they'd compile just fine and run pretty fast. I wrote a paint program (Paint!, image below) and a couple of plain (and simple) CGA games (Alien Invasion and BreakOut!, below) in Turbo Basic. In the end, Turbo Basic was still not fast enough and four colors weren't sufficient.
More importantly, there was one specific technical issue that simply could not be solved with any sort of BASIC. I wanted a four-way scrolling world, like those in nearly every commercial game of the time. I was so desperate to make this sort of game that I spent what little money I had on a commercial library (for BASIC) that supported scrolling through assembly-language routines. It turned out that the scrolling supported by the library was simple text scrolling through the BIOS routines, wholly inadequate for what I wanted. I was extremely disappointed and called the (apparently) single-person company for a refund. He refused. There was only possible solution - to learn C.
C had (and still has) a reputation of being extremely difficult, yet it was the language most commercial games were written in. I bought Mix C - a command line compiler/linker and nothing else - for $20. Coming from BASIC, it was a difficult environment. After floundering with that for months and maybe a year, I finally broke down and bought the Turbo C compiler for $150 and spent the next two weeks trying to build a working piece of software in C. Although I didn't fully understand pointers at the end of the two weeks, I wasn't far off from a basic working knowledge of C. Good tools can make all the difference. I still couldn't figure out scrolling until I received a letter from Herman Miller that brought everything together. From that Legacy I (shown above and below) in C finally came to be. It was a collaboration between Eric Lambert (who did the maps and much of the game design) and I. In truth, it wasn't a game in the end - the player is merely able to roam the world in full 16-color Tandy splendor. That was good enough for me because we now had a new computer with 256 (!) colors. It was time to start all over.