At the end of 1993 I was very, very excited about the newly launched Newton platform and the possibilities it represented. As a result I decided to pack up and leave the best job I've ever had. I had been a developer at the University of Michigan's Office of Instructional Technology for three years, and worked on a wide variety of educational multimedia software.

It was fascinating work: I got to bounce around between writing C, HyperCard, Toolbook, Visual BASIC, and other platforms. I had gotten some experience with video production and editing. I had written articles for the Information Technology newsletter. I had taught classes. I got to work directly with faculty members in the process of designing new instructional tools. I liked my co-workers. I had excellent benefits.

But I was not satisfied with the relatively low pay. It was also clear that there was not going to be an immediate opportunity to develop Newton software, and I wanted to jump into that world with both feet. It was a case of "the grass is always greener" and I had ambitions to do more and learn more, and was privately hoping to work up to what I considered the holy grail -- one day working for Apple itself.

Pharos Technology flew me down to their offices in Maineville, Ohio. I was young -- only 26 -- and had learned to drive relatively late in life, only getting my license two years earlier. Unaccustomed to city driving, I did not want to get lost navigating around the Cincinnati area in a rented car, so Pharos was very accommodating, and sent a driver carrying a sign with my name on it. He delivered me safely to my interview.

Pharos was a modestly sized consulting company, doing primarily Macintosh development on vertical market applications, such as control systems for satellite hardware for Hughes, and kiosk systems for Hallmark. They had an extremely nice "caves and commons" facility, with gorgeous meeting rooms and lots of windows. The atmosphere was sunny and friendly.

The project in question was the Infielder Crop Record system, developed for Monsanto, an ambitious application designed for farmers. This project was on a fast track and the team was concerned; internally, it was referred to as a "Wild West" project, because with so many unknowns and an accelerated schedule, it was not considered amenable to formal time estimation. A formal design still existed, though, with very extensive paper screen designs and descriptions of data flow through the system.

I had gotten up extremely early for the flight, and was quite nervous, so my interviews are a bit of a blur. I remember that I showed the Pharos engineers some HyperCard XCMD code written in C. There was a somewhat intimidating code review by a Macintosh programmer, but I knew C better than I did NewtonScript, so even though I had forgotten to print my header files, I convinced them I was worth hiring. By the time I got home that night, a job offer was waiting on my answering machine. A considerable raise from my University job was on the table, as well as assistance with moving expenses.

I had to move quickly, and left everything behind -- a house I was renting with my girlfriend, and most of my possessions -- to take one carload of personal effects down to Maineville, Ohio and set up an apartment near Pharos. I had a wok, a flimsy sling chair from Ikea, a phone, a futon, some clothes, and almost nothing else. I bought a TV and VCR and spent my evenings working out in the apartment complex gym or watching episodes of Mystery Science Theater 3000.

Although I had been hired as a software engineer, I was disappointed to find myself assigned to the project primarily as a tester. The project manager, Rick Armstrong, a plain-talking and pragmatic guy, was friendly enough: he provided me with a Mac Quadra to set up in my cubicle, then left me alone to, as he put it, "finish pissing on your stuff." Theresa Butt was my direct manager, the lead for the testing group, and also friendly and approachable. There did seem to be a bit of contention over exactly my boss was, though, and I would occasionally get a confusing lecture from one of the two for doing something the other one had asked me to do.

For the first few weeks, I spent most of my time running "test case scenarios" using FileMaker and recording the results, although my cubicle was right next to the engineers, not the testers. There was a virtual "wall" between testing and engineering, and when I complained about not having enough to do, my manager Rick Armstrong would talk about throwing a build "over the wall" for me to test. I found this very frustrating because I was at Pharos to program. The physical cubicle walls were low, and my desk was next to the engineers working on the program, but yet I was behind that wall!

As things went on, a few things quickly became clear.

The first was that the the complex application we were trying to build was severely taxing to the rather underpowered original MessagePad. Infielder was very complex, with screens containing dozens of interface elements. It was an extremely common occurrence to have the system just lock up hard while opening a view, leaving the user doodling on buttons instead of tapping them. (The "inker" task almost never died even when the NewtonScript application environment had long since crashed).

The ambitious Infielder GUI attempted to collect a lot of different types of data on a relatively tiny screen, resulting in a very cramped user experience. It was very difficult for me to use, even under good light, with sharp eyes. I had a hard time believing that any farmer trying to get real work done would tolerate such a frustrating and tiny user interface. Apple had consistently recommended that Newton applications should be relatively small and light in design, and this one would have been much more at home on a desktop computer. The design did do some things right, though -- it did not rely heavily on handwriting recognition, instead providing pop-up lists of choices for data entry whenever possible.

The second was that the development team of Ricardo Martinez and Scott Johnson did not seem up to the task, at least not at first. Development for the Newton required enormous attention to detail, and the Newton Toolkit was in a relatively primitive state. Ricardo had previously been a HyperCard developer, and Scott a Macintosh developer (he had developed a popular shareware extension), but neither seemed totally comfortable in NewtonScript. A background writing Lisp or Scheme code would have been much more valuable, but NewtonScript can initially appear to be more like Pascal, BASIC, or even HyperTalk. Mike Engber at Newton DTS was a bit disparaging about the engineering team, asking me once "is Ricardo the top programmer at Pharos?" Unfortunately, I suspected that Ricardo had been given a task for which he was insufficiently prepared.

In testing, the same kinds of bugs popped up repeatedly. The code was a mess, with a large number of redundant functions, so identical bugs often had to be fixed in multiple places. The engineers slogged through the slow Newton Toolkit build process. I thought I could do better. Scott privately told me "you probably know more about Newton programming than I do." I was eager and itching to prove myself. My self-confidence was not necessarily fully justified -- as a C programmer, I had only minimal practice with languages like Scheme, and several of the concepts behind NewtonScript were beyond my experience. But then, where would a software engineer be without hubris?

I was able to contribute occasionally. Prior to a deadline, the team wound up working through a weekend, staying up several nights running, catching occasional catnaps on the floor of our cubicles, and living on pizza. In order to test the application, it was necessary to have soups populated with the data to fill the interface's configurable pop-ups. The project code name was "FarmBoy," with an internal logo consisting of a Newton carrying a stylus and wearing a cap. I quickly put together a small application I called "FrameBoy," and drew a cute icon, which built up the relatively complex soup entries required. I was quite proud of FrameBoy. So for a time I was coding alongside the engineering team, and FrameBoy saw at least some use.

After Ann Arbor, life in the bedroom suburb of Maineville, Ohio left me bored out of my skull. There seemed to be no bookstores within a short drive, no restaurants other than bland chain establishments, and nothing to do in the evenings whatsoever. The only advantage I could discern was that groceries were much cheaper than they were in Ann Arbor. I was intimidated by city driving and did not feel up to trying to explore downtown Cincinnati itself, especially without someone else to go with me for moral support. Some evenings I would go running, but quickly was stymied by the near-complete lack of sidewalks in the neighborhood. I sometimes found myself having to run in drainage ditches, and one one occasion was threatened by an angry homeowner when I cut through his backyard. I tried exploring the surrounding communities, but it wasn't that much fun to do alone.

I had purchased a PowerBook Duo, and in the absense of a social life I spent my time off from work programming as well, sitting in my little sling chair in the otherwise empty apartment working on my PowerBook, writing programs and articles for Steve Mann's PIE Developer magazine.

I did not have enough to do at work. I was not considered an engineer by the engineering team. Ricardo and Scott along with manager Rick Armstrong were sent to a Newton developer event in early 1994, and I was not. So not only was I not doing Newton development, but the job was preventing me from attending a developer event I would otherwise have paid for out of my own pocket, as I had the first Newton platform conference. This was not what I had signed up for. I heard through the grapevine that a limited edition of newly minted clear MessagePad 110s were available to developers at the conference. I wanted one badly, so by e-mail I asked if someone from the Pharos contingent would buy me one, and I would happily pay them back upon their return. I received a return message from Rick Armstrong telling me that they had acquired a clear Newton with my name on it.

Ricardo had bought it for me. But when he returned, he apparently realized that instead of reselling it to me, he could sell it elsewhere and turn a profit. He told me he that he had gotten a better offer, which I would have to beat, and named a considerably inflated price. I refused to pay, and so lost my clear Newton. I did not have much more to say to Ricardo during the remainder of my time at Pharos. Almost fifteen years later, I am still galled when he recalled that he considered this behavior a reasonable way to do a favor for a fellow team member.

I worked on several other bits and pieces at Pharos: we developed a demo using Wayfarer, an early "middleware" system for communication between the Newton and cross-platform desktop applications. I was tasked with developing a demo on Windows for Pen, using Visual BASIC with pen extensions, but although the demo worked, it was clear that Windows for Pen was a true train wreck, and unsurprisingly Microsoft rapidly dropped it.

In the evenings I continued to work on my articles. I was one of the first developers outside Apple -- perhaps the first -- to use the infrared toolkit. This became another short article for PIE Developer. While I still didn't get out much, I finally found that I got along pretty well with a co-worker named Jeff Knee, who also played guitar, and we did a few things on weekends; it was much more fun to explore the area with a friend.

I was inexperienced with organizational politics. There are times when it is wise to just hunker down and look busy. Perhaps I spent too much time in the Pharos lunch room playing with my little Newton program to change channels on the TV. The environment became tense. We heard weekly updates from Bob Beech, the CEO of Pharos, about the increasingly grim money situation. At one point I was pulled into a manager's cubicle and given a whispered pep talk reassuring me that because of my Newton skills, I probably would not be on the layoff list. But shortly after a remarkable total solar eclipse that brought the entire Pharos team out of the building to experience the weird light, I was culled from the herd.

One morning, the entire company received notification by e-mail whether they would be staying or going. I was one of the ones going. Many of the newer employees got the axe that day. I don't recall the numbers, but it was perhaps fifteen or twenty of the company's forty or so employees. Some got a couple of days notice, but I was asked to clear out immediately.

I had loaded up my cubicle with heavy textbooks on algorithms, Scheme, and C. While the "Pharosians going forward" sat in the glass-walled conference room listening to a pep talk by CEO Bob Beech, I hauled load after load of my books past that panopticon, in what seemed like an endless number of trips up and down the stairs to my rusting Plymouth Reliant in the parking lot. I felt like a bug under a microscope, although in retrospect I doubt if anyone was paying much attention to the long-haired new guy who had lasted only three months.

I had my brief pep talk from HR. I was only in Ohio a few days later, long enough to attend a get-together for the "Pharosians not going forward," where I drank a very large beverage involving citron vodka, said some goodbyes, and stumbled back to my apartment in a bleak mood.

Apparently some layoffs had been premature, and one or two of my fellow culls were frantically re-hired on a temporary basis to complete some work that was actually bringing in income. But the decision in my case was final. Three months of the higher salary and a few weeks of severance pay didn't go very far when pitted against the laptop I had purchased and the prospect of unemployment. It was a mournful time as I packed up and headed back to Michigan and back to my girlfriend, unsure of my next move.