After being fired from Pharos technologies, I returned to Michigan. and moved back into the house I was renting with my girlfriend. There were some money issues terminating the lease in Ohio, but nothing too serious.
I had remained in touch with a man I met at the first Newton development conference, David Fry. David Fry was the proprietor of a small consulting company known as Gulfoss software, named after a waterfall in Iceland. With financial backing from his family's printing business, David was forming a startup, Fry Multimedia. I wound up coming in to Fry Multimedia as employee number 4. David was like many startup founders: smart, funny, opinionated, and very plain spoken. Working for a startup can be very difficult but you may also find yourself becoming an "adrenalin junkie." The opportunity to finish very challenging projects with few resources and crazy deadlines can be very satisfying, but only if you don't crash and burn first.
The company had a small suite in an office building on the far edge of Ann Arbor, only a twenty-minute commute away. My office was in what was formerly a very fancy lawyer's suite, complete with a wet bar, which unfortunately was wasted on us.
Fry was bidding for development work on a variety of projects. My co-worker Antonio Antiochia had done game development while still a teenager, and was assigned to develop a demo game for HBO, based on the "Tales from the Crypt" television show. I developed a simple scriptable multimedia browser in C++ with QuickTime support, for a proposal to Scientific American magazine. The CD-R was quite new at the time -- I remember we spent several thousand dollars for a CD burner and made a lot of coasters -- but Fry Multimedia was trying to come up with new uses for this medium. But David Fry was also a fan of the Newton, and wanted Fry to be an early and active Newton developer. So while I continued to spend my evenings working on bits and pieces of Newton code and writing, I had some paid development work to do, putting together proposals and demonstration code for Newton development projects.
One project was a complicated one in which Fry would act as a subcontractor for another contractor, developing a tool for a carpet store chain. I worked out a paper prototype of a user interface whereby a contractor could measure a set of rooms, edit any irregularities in shape or area with the Newton pen, and quickly come up with an estimate for selling square feet of carpet. The carpet layers had previously been using a ruggedized laptop solution, but these were expensive and finicky.
We had met with the main contractor, and things seemed promising. The people involved seemed satisfied with the ideas I came up with in the paper prototype, which included ways to quickly edit room shapes. Part of the project effort would be interfacing with a laser range-finder, a serial device.
One of the early questions from the customer was whether the Newton interface could be used effectively to manage diagrams larger than the screen, by taking advantage of scrolling and zooming. The Newton's built-in applications did not generally scroll smoothly, but in screen-sized chunks. As a proof-of-concept, I adapted a piece of Apple's sample scrolling code to scroll a sample floor plan.
The next step was a meeting at the customer site. The carpet company's offices were busy and full of staff, but it couldn't have looked any more different than the usual shiny and clean spaces used by software companies. If the underlying message at Pharos had been that they were very proud of their customized facilities, the message at the carpet company was "we don't spend one cent more than we absolutely have to." Clerks were crammed in at tiny desks working on ancient PCs, and everything was exposed plywood and ugly old carpet with a dingy, worn-out look.
The meeting did not go well. We had what I thought was a fairly conservative proposal, including a couple of development milestones, and a testing phase including field trials. Our potential clients started on the offensive, telling us we were asking for an order of magnitude too much money. They had called Apple asking about whether the contractor was a registered developer, and found that they were not, and thus we had to explain that Fry, the subcontractor, was the registered developer. They did not seem happy with that. When I showed our little proof-of-concept, they took the position that we shouldn't need to charge them much at all, because the program looked pretty much done.
As flattering as that was, there was nothing at all behind the demo. I lost my cool and burst out in frustration "there's a big difference between a picture of a program and a real program!" I then started feeling something banging against my leg. My boss was kicking me under the table! The message was "shut up and let me do the talking." I may have been young and dense and not really suited for this kind of hard-nosed negotiating, but I eventually got the idea and let Dave do the talking.
The tone of the meeting did not improve, especially when one of the managers claimed that he could hire his brother-in-law to write the program in a week or two. I think I wound up saying that they were certainly welcome to try working with the brother-in-law, and if that didn't work out well for them, they could get back in touch with us. The meeting ended.
We later heard that they were claiming that they had contacted someone at Apple to see if our quote was "in the ballpark," and that Apple had told them we were asking far too much and our schedule was far too long. I never did find out whether that was another negotiating tactic, or if it was actually true. Anyway, that seemed to be the end of Newton development at Fry Multimedia.
I went on to spend a few months working extremely intensively with Antonio on a search engine for a CD-ROM database product. We worked on a very aggressive schedule, working the stereotypical crazy hours, sometimes 70 to 90 hours a week, staying up multiple nights running, grinding my teeth until my jaw popped and I needed a bite splint to sleep with. I'm still
proud of the code I produced during that push, and the product was finished to beta quality. But a direct competitor got a similar product out first, and ultimately our search engine did not become a product either.
Fortunately, Fry Multimedia was able to make money with a new thing called the World Wide Web. I pretty quickly got tired of writing HTML by hand, though. But I had received an e-mail that someone at the University of Michigan was looking for a Newton developer. That was still what I wanted to do, so I left web development behind and took a job with the nascent Health Media Research Lab, willing to take a pay cut in exchange for lower stress and the chance to developing something truly useful for the Newton. So it was time for yet another change!