I transferred into the Network Engineering group in 1989. At last I was getting back to my technical roots. Some of my publications skills proved useful in the new position. This aspect solved certain problems even as it created others.
The only real qualifications I had for the job was that I knew a fair amount about Unix and that I was reputed to be a quick learner. Our Sperry mainframe machine had gone the way of the dinosaurs, but we had a cluster of large minicomputers, DEC Vaxes running VMS. Most of the staff used the VMS machines only for email services, however, for 90% of the real work was being carried out on desktop machines, whether DOS-based or Macintoshes. (Windows was not yet the Leviathan it has since become.)
Somewhat against the usual trend of using Macs for such work, I had been doing almost all our documentation on DOS machines. With my mainframe background, I was more than comfortable using the command-line interface. But one of the Vaxes was running Unix (well, Ultrix, which was a sort-of Unix), and because I could have complete access to Usenet news through that machine, I got busy learning everything I needed to know. At the time, we had very little in-house Unix expertise.
But I didn't have much experience in the real issues of networking, beyond knowing such basic things as the difference between a bridge and a router. I had a reputation as a good software tool-maker and would immediately be able to apply those skills in the new group.
In addition, the networking group had a number of publications projects of its own, and it was correctly assumed I could get those polished off fairly quickly, however joylessly. The last phase of the previous job had been so painful that I intensely disliked any reminder of that period.
A potential downside to any job, I would say, is that you can get stuck in a rut through no fault of your own. Sometimes things just work out a certain way, and the consequences can lead to what to me seemed very much like a trap, even when it resulted from doing a good job with things that really needed doing. I felt appreciated but unfulfilled as a result.
At the time our group was responsible for operating the statewide regional network, WiscNet, under a facilities management contract. WiscNet was a fast-growing concern and there were a number of reasons for me to be one of the principal staff people associated with the network. That too involved some publications work, rather to my annoyance. I had long since gotten a Unix box (NeXTstep, actually) for my desk, and I also bought a NeXT machine on the used market for use at home.
WiscNet needed a newsletter (or thought it did; only a couple issues were ever produced), and who better to create it than your scribe? I had been fiddling with the only document system worthy of the name running on NeXTs, I thought: FrameMaker. So I designed the newsletter, wrote nearly all the copy, did the art, and everything else. On a lark, I submitted it to a national contest Frame Technologies was sponsoring and got a rather nice plaque for the honorable mention it was awarded. Not great, but not bad, either, for my first FrameMaker project.
WiscNet proved to be the first of several detours diverting my technical path away from network engineering proper. In the summer of 1993 the UW decided to create a "free" dial-in pool for all students. To be sure, there was a real network engineering aspect to that idea. But I was tapped to be the project leader, and the really large effort was not setting up modems and terminal servers, but assembling and testing a software package to support dial-in services.
This was a very high-profile project because the University administration, including student representatives, had actively lobbied the State legislature and the Governor's office to enact a tuition surcharge to fund this technology initiative. I asked for and got a firm commitment for unlimited staff resources because our time scale was so short: less than two months.
Our team had 17 people on it, and in 57 days we had the product in production, up and running on the first day of classes, complete with trained help-desk support, documentation, software for Macs and PCs, and the whole nine yards. But it had necessitated 80- to 100-hour weeks for many of us and we were pretty fried.
Unfortunately for me, and though others worked harder than I did, this project gave me a great rep for project management, and I was soon asked to take on -- this time without a large team or unlimited resources -- the task of creating a similar service for WiscNet, so we could take advantage of NSF funding being made available for such things through NSF's Rural Datafication initiative.
In all, nearly two years of my life went away while I was being Mr. Dial Access. There have been several stages of expanding facilities and services for both UW-Madison and WiscNet dial pools, but in early summer 1995, we finally managed to move everything from development mode to production mode, which means others would be taking care of things.
An important colleague left in 1995 to take another job (paying much more, of course), and as a result I became Mr. ISDN -- dial business of another sort. Ameritech, the major phone company in the five-state midwestern region, ran a pilot project with us and others to help develop its ISDN product services. One result was that I got a direct connection by ISDN from my house to the campus, with the University footing the bill.
In July, 1997, after 41 years of service at the University, I retired. Our retirement annuity plan makes it possible to return to work part-time, without a reduction in benefits, if there is a need for one's particular talents, so after a break of about six weeks, I went back. I was worried that the retirement annuity would not provide enough income to sustain me without a major reduction in lifestyle-related expenditures, so I took a 40%-time project appointment, continuing the same projects I'd worked on before retirement.
Hindsight is a skillful informant, and now it's easy for me to see what a mistake that appointment (and the second one that followed) was. The hope that at last I would be able to get my hands dirty with real network engineering faded very quickly. Giving up was the low-energy course; there was too little to be gained by trying to buck the system.
As it turned out, I was an idiot to have worried about income: I have more than enough and I've been much happier being truly retired.