After I lost my job at ICT, it took quite some time to find another. The job market in Ann Arbor in 1988 wasn't that great. For context, this was the end of the Reagan years and the beginning of the George Bush Sr. years, with the S&L crisis in full swing. It was also when Joe Biden first appeared on many people's radar. I ended up working night shift at a Kinko's copy center, which was definitely a step down from my ICT job but not too terrible by the standard of other jobs I'd had. I was also pretty active in the M-Net community at this time. An M-Net friend who had moved to Massachusetts suggested that I send him a resume to shop around there. I didn't really expect much, and wasn't sure I wanted to move to another state, but I sent him one. As it turns out, I did get a hit at a company called Technology Concepts International. Yes, my first two industry employers both had names that could just as well have been for vacuum-cleaner companies.
It would be hard to overstate how wrenching this move was for me. I had to leave what little family I had, my first significant romantic relationship, the M-Net community and my few (but closer for that) friends. I was going from places I knew and reached on foot to places that were strange and driving everywhere. But in 1989 I bought my first car - a Ford Escort that barely lasted longer than the trip itself - packed it with my few belongings, and headed east.
My first job was actually in quality assurance, not development. I found it oddly satisfying at the other end of my career to have come almost full circle and be focused on testing again. When I arrived, my boss dropped off a bunch of reading material to get me started, then (as far as I can tell) went upstairs to quit. Such a vote of confidence. A few months later, my second boss traded me to development in exchange for a sandwich. If I recall, it was at the local Restaurant 99 ("the nines") which was our most frequent hangout after work.
TCI had been founded by a couple of DEC alumni who had been principals in the development of DECnet - an alternative to the ubiquitous TCP/IP network stack we all use today. (It makes me slightly sad that I have to explain that BTW. Part of the reason I'm writing all this down is that our history is too easily forgotten.) Already part of Bell Atlantic by then, TCI's basic charter was to provide DECnet for all the platforms that DEC didn't - Macs, DOS (still a thing back then), both System V and BSD UNIX on VAXes, and a bewildering array of embedded devices. I remember both pSOS and VRTX being in the mix. I also remember my first trip to New York City to work on a Stratus VOS port. Here are a few things I learned on that job.
- DEC folks knew how to design network protocols. LAT (a terminal protocol that was not technically part of DECnet but closely related) had scale and latency-tolerance properties at least a decade beyond contemporary telnet or rsh. The protocols I worked on the most all had robust feature-negotiation handshakes based on "extensible bitmaps" that I often wished for when I worked on similar things in the TCP/IP world much later.
- They also knew how to write stuff down clearly. Far better than the internet RFCs of the time. Throughout the rest of my career, I considered experience at DEC during that time as a strong positive signal for anyone I interviewed.
- They also had flaws. They cheated. Despite all that fancy feature negotiation, some things just wouldn't work right, or wouldn't work optimally, unless you identified yourself as another VMS system. So that's what we did, no matter what platform we were actually running on. I saw this same kind of BS from Sun later, in both TFTP and NFS, but I'm getting ahead of myself.
- Some terminals had a BNC connector on the back that was physically the same as that used for (thin) Ethernet, but was not electrically the same. Learned that the hard way when one of our support guys plugged his terminal into the company network.
- Speaking of terminals, I was the envy of my peers when I got one with 40 lines and two serial ports so I could have two sessions at once. Such luxury.
- Complexity shows up in surprising places. One of the hardest problems I had to solve at TCI wasn't some tricky low-level protocol thing but the simple act of translating file names between different operating systems. Granted, things were pretty complicated back then. We had VMS, UNIX, DOS and Mac all with different formats, separators, and allowed characters. The test suite for this one seemingly trivial part ended up being huge.
The other thing I remember from TCI was the social environment. The CommUnity team, of which I was a part, was tight. We spent quite a bit of time together socially - "beat the Lakers" parties, my first whale watch, a strip club, hiking, and so on. I'm still in touch with people from there, more than for most of my more recent jobs. I now know that was a function mostly of our age and personal context rather than anything about the place or job, but it's still special.