I'm going to jump on the bandwagon...
I wrote my first program on paper by following some snippets of basic code I found in a sidebar of my math text book. I remember writing multi-page basic programs on paper long before I actually had a computer I could use to run them.
I first got paid to write code for an Ohio Scientific Challenger III. I developed some multi-level sales tracking software for a short lived company. I thought I had invented hierarchal storage structures and recursion (then I took CS101 in school).
Other early jobs were done on a Lanier something or another, a Kaypro II, and, of course, an original IBM PC.
I hacked CP/M on the Kaypro to make a disk driver that talked out the serial port to an app running on an HP3000 that maintained a file that was formatted like a big floppy disk. We actually supported multiple users simultaneously. Teachers used this to distribute assignments and collect homework. We also redirected the parallel port for printing.
On the PC, we wrote a TSR that did the disk/printer redirection as well. We also added a feature that let the school distribute (on floppy disks) more copies of software to students then they actually owned. We took 15 bytes of code out of the startup sequence of each app and stored it in a database on the HP3000. We replaced the 15 bytes with a call to our TSR that would check the number of people currently running the program and provide the missing code if all was OK. We even had an offline mode. Some PCs had ROM chips that we could pull the code from in case the HP wasn't available. Funny how I still remember it was exactly 15 bytes.
BTW: http://www.obsoletecomputermuseum.org/ has an RSS feed.