In the fall of 1994, I was a Freshman at Wesleyan University in Middletown, CT. I got to school and promptly went looking for the student computing group so that I could apply to get a Unix shell account (1). Apparently, I said the wrong words, because I found myself filling out an application for a job working at Wesleyan's computing helpdesk. At the time, the helpdesk had a VT220 terminal with a mail client that stayed open all the time, so that whoever was on duty could respond to incoming support requests. Lots of things fell through the cracks, but we did ok.
That summer, I worked for the university improving helpdesk and residential networking software infrastructure. As part of that project, I decided that the helpdesk actually needed their own server. I managed to scrounge a 486DX25. I spent a whole afternoon and evening performing my first-ever Linux installation on that box. 25 floppies later, it was 3am and I had a sort of functional server. Except I'd messed up the installation. By 8am, I had a working Linux box.
A day or two later, I finally got around to installing a helpdesk package that I'd heard great things about: req, by Remy Evard at Northeastern University. Starting a week or two later, Wesleyan's helpdesk started using req to track support requests. At the time, the helpdesk was mostly staffed by Unix-savvy hacker/sysadmin types, but we had a few Computer Science majors too. Not all of the CS types were so comfortable with the command line tools that req presented and only one of the machines in the helpdesk office could run X11 to use req's Tk interface. But, the Mac running System 7 and the new Windows 95 box could run Netscape 1.1. So I embarked on my first actual programming project: building a web interface to req.
I spent the summer and the next fall semester teaching myself enough perl to put together a CGI interface to req. Eventually, I worked up the courage to tell the world that I'd been doing so:
On 12 February 1996, I released the first public version of WebReq (2):
An employer eventually forced me to throw out the "req" portions of WebReq and build my own ticketing system underneath, but 90% of the code in WebReq became part of RT 0.9. Which became RT 1.0. Which was gutted and reinvented as RT 2.0 in the summer of 2001 when I quit my day job to work on RT full time.
Nine years ago this Saturday, I released the very first version of what would become RT. I never, for a moment, dreamed that I would still be working on ticketing today. And not in my wildest dreams did I imagine that I'd earn my living writing software and giving it away. And the fact that I get to pay others to build and support free software -- that still blows my mind.
Thanks, everybody, for the first nine amazing years,
8 February 2005
(1) At the time, Wesleyan was a VMS shop. Shell accounts were provided by "The Student Net," who had gotten their hands on some old NCR towers and a couple of Apollo DN2500 workstations that had been pressed into service as servers.
(2) You'll note that even before the boom really started, I was fond of that horrible InterCapped style.