20 Years of Computer SoftwareI've done a lot with computers in the last 20 or so years. I think I know more about the details of how computers work than I otherwise would because of the way I started to learn in those first years. I couldn't learn by trial and error, because I didn't have access to real computers.
1976: BASIC, 6800 assembly, and binary math -- in books.My father, who still works at Motorola, gave me a Motorola book named M6800 Microprocessor Application Manual. That book describes how to make a cash register with a Motorola 6800 microprocessor, assuming you already know what a microprocessor is, how a 6800 works, what a "branch instruction" is, and other background stuff. The book is large and thick with about 500 pages, but only about 20 pages are devoted to details of the 6800. I read this book about a hundred times, every night before I went to sleep, figuring I would understand it eventually.
An older cousin heard I was interested in computers, so he gave me a book by IBM that listed all the BASIC commands. The appendix of the book described the computer that ran this dialect of BASIC -- according to the appendix, it was available in an APL version; sounded interesting but I had no idea what A P L stood for. It was pretty easy to start writing BASIC programs once I read this book. I didn't have a computer to run them on, but I had a lot of yellow lined paper, and a few red pencils, so I wrote the programs down.
At school, I found a cool self-paced instruction book that taught about binary, octal, and hexadecimal numbers. It's great how 3 bits turn into an octal digit or 4 bits into a hexadecimal digit, and I was lucky to find this book. I have no idea why it was in the library. If anyone knows where to a find a copy of this book, please let me know. I'll post some pages from it if I ever find it.
The programs I wrote during this period were never run. Also, I got everything wrong in them. I still have some of those pads of yellow paper filled with incorrect BASIC programs.
1977: Tried programs on a real computer.I got permission to send cards with a BASIC program on them to a computer at the local high school. The cards were like punch cards, but instead of typing on a card-punch machine, you filled in boxes with pencil. My first program was a sabotage program. I knew about a very large number called a "googleplex" and I decided to use this knowledge to make the computer do something funny. I sent in a very short program that went like this:
10 PRINT 10**(10**100) 20 END
I imagined the page after page of zeros pouring out of that little HP computer's thermal printer. I had never seen what the computer looked like, but I could picture the havoc at the high school when it went out of control.
I got back my two cards, and a printout that was the result of running my program:
OVERFLOWRude awakening. Floating point. The BASIC manual didn't say anything about this.
1978: Little computers in my neighborhood.Computers started popping up all over the place.
The local mall had a store called Mr. Calculator. They had something there called a Commodore Pet. It was a cool looking computer that you could program with BASIC. The guy running the store wouldn't let me touch it though; there were always lots of adults who wanted to play with it. I kept hanging around, reading the copies of Kilobaud and Byte that were in a little magazine rack at the side of the store. Finally, the guy behind the counter said that I could come back on a weekday.
I got my mom to drive me to the mall on a Wednesday afternoon. When I got to Mr. Calculator, the guy behind the counter couldn't remember me from the weekend and wouldn't let me touch the Pet. I was pretty disappointed, went and got my mom, and told her it was time to leave. Before we left, I think she went and chewed out that calculator salesman, but I was waiting for her in the car so I didn't hear any of it.
Other computers, Radio Shack TRS-80 machines, started showing up. You could use them in the Radio Shack stores. My school bought two of them. They didn't know what to do with them, so they put them into a little room in the library. Some guy visited our little room to talk to kids who wanted to know about computers. He brought one of those cool Silenttype terminals with the acoustic coupler modem and thermal printer. He showed us this Star Trek game. It used up a lot of thermal paper, but it was also the coolest thing I had ever seen.
I turned 13, and my parents got me a computer as a Bar Mitzvah gift. My grandparents got me a power supply for it. It wasn't a Commodore Pet, or an Apple II, or a Radio Shack TRS-80, or even a KIM-1. It was something much better: the Motorola MEK6800D2 Evaluation Kit. This was a computer with 256 bytes of RAM, a little hex keypad, and six 7-segment LEDs, and a 300 baud interface so you could store programs on audio cassette tapes. You had to solder it together, but all the components were included. My dad helped me build it. When we finished, it didn't work, so he took it to the office. Some guy there used some kind of fancy diagnostic tool. He said we had a solder bridge between data bus lines D2 and D3. We didn't know how to find that problem with my dad's oscilloscope and my logic probe, and hours of poring over all our solder joints didn't find it either, so I was lucky that my dad could call in the pros.
There were some cool programs for that D2 kit. One of them was a lunar lander program that fit in the 256 bytes. Cool.
I wrote an interpreter for a little language that let you make messages flash on the LEDs. I remember a program for that that made one LED segment dance around the display, and another that made the LEDs flash the letters PLEA SE alternating with HELP US. There was a space between the A and the S in the word please, but I thought the overall effect was pretty convincing. I told my younger brothers that there were tiny insects in the computer that wanted to get out.
1979-1983: Personal computers arriveThe next four years, there were computers popping up everywhere. My high school bought a bunch of TRS-80 computers, and a couple of Apple II computers; they also continued to use that cool HP machine with a line of 32 alphanumeric LEDs, a thermal printer, a digital cassette drive, and a 4-pen plotter. There were also disk drives on one TRS-80 and one Apple II.
I'm a pretty competitive guy, and I decided that I was going to learn more than anyone else in the school about these machines. My friend, Nevin Liber, led me on a merry chase, though. He was also learning about these machines at a frantic pace. When I first met him, I knew a little bit more because of the books I had read. I desperately wanted to remain ahead of him. I found a winning strategy. I would casually question Nevin about what he was interested in. Then, I would quickly study up, so that by the time he wanted to find out about the stuff, I was already an expert. I learned Z80 assembly language that way. Nevin said something like "I think maybe I should find out how to write assembly language for this TRS-80." I went out and found a book on Z80 and read it that night. The next day I said something like "I know Z80, do you want me to teach you?"
We wrote tons of computer programs in those next few years. Also, we played lots of games on Apple II machines our friends owned. That high-resolution graphics mode made the Apple II much better for games. Our richer friends who had Apple II machines at home became enterprising game software pirates.
My dad bought a computer for us to use at home. It was a Radio Shack TRS-80C Color Computer. This was a pretty cool machine with a Motorola 6809 processor in it. I wrote some good programs for it, including a version of Star Trek. My dad said that I should call it "Adler Trek," so I did.
1984: MacintoshDuring my first few months at college, during 1983, I got excited about two new computers: the Corvus Concept and the Apple Lisa. Both of them had black and white bit map displays. They didn't use character generators. On the Concept, they did it so you could rotate the screen and go from portrait mode for the word processor to landscape mode for the spreadsheet. On the Lisa, they did something more radical. There was this whole thing with a mouse and icons and windows. I went to a local store and played around with a Lisa computer. It was the coolest thing I had ever seen, but it was little boring after a while. Also, it was so easy to use it was scary. It seemed like you could really easily screw yourself by throwing away the application program icons or the stationery icons. After I saw these two machines, I knew that I needed to program a computer with a bitmap display.
Early 1984 my friend David Conely showed me a copy of a magazine called MacWorld. It was filled with pictures of screens that had bitmapped displays on them. It was like the Lisa only it looked better -- not as ornate. Steve Jobs talked about how great the computer was. There were nice looking ads for it. Also, there were pages filled with the pictures of the people who made the computer at the back. I memorized the names and faces and even the job titles: Steve Jobs, Andy Hertzfeld, Susan Kare, Bill Atkinson, Steve Capps, Mike Boich, Joanna Hoffman, Martin Haeberli, and a ton of others.
I was seized with an immediate lust for a Macintosh computer. One showed up at a local computer store, and I went there regularly to play with it. My friends and I read about something called the Apple University Consortium. For us it was, "You can get a Macintosh at a discount!" Unfortunately, they had a complete list of schools which did not include our school, The University of Texas at Austin.
This was the first time in my entire life that I wanted something that cost more than the money I had on hand. It would take more than $2000 to buy a Macintosh. Then, the good news came in: Apple had added The University of Texas to its program -- we were the last school to be added before the program began. Unfortunately, it would still cost about $1500 for me to get a Macintosh at the low school price. At that price, I wouldn't get MacWrite or MacPaint, but I figured I could probably snag an illicit copy of those.
I heard that the university ACM chapter had a copy of Inside Macintosh, the manual that told how to program a Macintosh. Since I was curious, I went to borrow their copy. They let me, even though I wasn't a member. That was probably a mistake on their part. I was extremely interested in the calls described in the 3-ring binder, so I took out the summary of calls from each chapter and returned the manual to them -- minus the summary pages. I studied those lists of calls over and over, trying to see through the names to the ideas of the programmers who invented them. The graphics system, QuickDraw, seemed to have some really great ideas in it. There was this whole idea about using coordinates for the points between the pixels instead of the pixels that seemed fascinating and brilliant.
I had to be sure about getting copies of those applications, so I bought a 3 1/2" disk from a local computer store that carried an HP computer that used them. I took it to the store that had Macintosh so I could copy MacWrite and MacPaint onto it. I was nervous and didn't know how to format a disk for Macintosh. I puzzled over this for a minute or two, and then just decided to put the disk in. The computer asked me if I wanted to format the disk! This was much better than the other computers. I could tell that the people who made this thing were thinking about how to make computers better; not just doing the obvious stuff, but paying attention to the details. It was easy to copy the programs and take them home. June 1984 I got my own Macintosh.
1985: Schoolwork suffers.During the summer of 1984, I went to visit a local software company, called TMQ Software, hoping to get some cool Macintosh software. I heard that they had a couple of Lisa computers, which is what you needed to program for a Macintosh back then. There were a bunch of programmers at TMQ and they had a nice-looking office. I struck up a conversation with one of the programmers, a 16-year-old kid named Waldemar Horwat. I didn't notice the people staring at us. It turns out that very few people were able to keep up with Waldemar, a brilliant young high-school student, originally from Poland, who had world-class skills in mathematics and computer programming. My conversation with him was enough to convince them to offer me a summer job. They offered $4 an hour, and that sounded like a good deal to me -- mainly because I was an idiot.
Waldemar was working on a debugger for the Macintosh that would use the mouse. Some other programmers at TMQ were working on an adventure game for the Macintosh. They asked me to write graphics compression for the game, and I started working on it. I shared an office with Waldemar; a big room with two small desks, two Lisa computers, and a lot of empty space. While I compiled Pascal programs to do the compression and wrote some 68000 assembly to decompress, Waldemar was writing his debugger, WHMON, in assembly. He didn't try to assemble it until he had written the whole thing.
Later that summer, they moved me into the same office with the other programmers working on the adventure game. Until that point, I was the only one that actually put any code into the computer. The other programmers were talking about the design. Once I was in the same room with them, I found it hard to limit my involvement to the graphics compression. By the end of the summer, I was an equal member of the 3-person programming team that went on to write the film-noir-style graphic adventure game "Deja Vu: A Nightmare Comes True" and the adventure game development system that became the basis for 4 or 5 more adventure games, including a horror game called "Uninvited," a sword-and-sorcery game called "ShadowGate." Waldemar's debugger was released later that year, renamed TMON after the company instead of the programmer.
I had a lot of fun working for TMQ Software, but it was really bad for my schoolwork. They sent the Lisa to school with me so I could use it to program. The two other programmers were also college students, one at Northern Illinois University, another at Stanford University, and me at The University of Texas at Austin. We racked up huge phone bills that the company paid working on the program and sending files back and forth across the country. They also gave us $50 a week while we were at school "no strings attached." We didn't have to work to get the $50, but it was pretty tempting. I didn't go to class much -- I worked a lot.
1986-1987: Macintosh communityDuring the next few years, I did lots of Macintosh programming. I got protoypes of the Macintosh SE and Macintosh II in my apartment in Texas, although Apple thought they were in TMQ's offices in Chicago. I spent lots of time on Macintosh newsgroups and I went to every MacWorld Expo convention. There was a great community of Macintosh programmers in Austin and a real sense of community with the other good programmers working throughout the rest of the country too.
After 4 years at UT, I realized that I wasn't going to graduate if I never went to any classes, so I went back to Chicago to work for TMQ Software, now renamed ICOM Simulations (don't ask). My parents were worried, but they let me move back into the house. I worked on a cool project back at ICOM: the debugger portion of a Borland product that never shipped -- Turbo Pascal for Macintosh 2.0. Many other people in the company were working on the adventure games that we had created back in 1985, so I was in demand to answer technical questions.
At the MacWorld Expo that summer, something really great happened. I was there to demonstrate the TMON debugger. Some folks from Apple's Macintosh Developer Technical Support asked me if I wanted a job there. I knew right away that I had to take the job, but it took a few months to work out the details. I felt a little guilty about leaving ICOM after only working full time less than a year, but I couldn't resist moving to California. A month or two later I was living in Sunnyvale, California and working for Apple.
1987-1991: Apple ComputerI went to company orientation for Apple. I recognized one of the people in our orientation group. It was Steve Capps. He told me he preferred to work on his own, but he was returning to Apple because he wanted to make a new computer, and you couldn't do that on your own. I soon learned that he was joining Steve Sakoman's Newton group; they were a small group making a handwriting-based computer that was going to be shaped like a book or a piece of paper, something like Alan Kay's DynaBook concept from his research at Xerox years before.
Working for Apple Computer was incredible. Macintosh Developer Technical Support was one of the best places in the company to work. We worked hard, but we also had a lot of fun. We christened the little dog-like thing in the Macintosh printer documents the "dog-cow." We did that thing that cliques always do: invented our own lingo with phrases like "he's in the weeds," "she's grazing off the cliff," and "don't forget the Fred rule." Some of our little jokes became famous throughout the developer community, which is not surprising, since our job was to communicate with them.
I had a great time. I edited and revised all of the Macintosh Technical Notes. I wrote a few new ones. I started the Macintosh Sample Code series by writing a simple sample application that is still used by Technical Support today.
Que es mas macho: Pink Finder or Blue Finder?
Then, I got an even better job offer. Jim Friedlander, who had been my hiring manager in Technical Support, told me that there was a job opening to work on the Macintosh Finder in the system software group. Working on the Finder sounded like the best job in the world to me. The Finder was the most important program on the Macintosh; the one that creates the heart of the user interface -- the icons and windows that you use to start applications and organize documents.
I jumped at the chance. After my interview, the manager told me that I was better suited to another job. There was new system software being developed at Apple, called Pink; the existing system software was now called Blue. My new manager, Gene Pope, wanted me to be the first programmer of a new Pink Finder team. I wanted to be the project leader of the existing Blue Finder team. The Pink Finder sounded good too, though. Writing the new Finder from the ground up would be fun, and the other programmers working on Pink were some of my heroes.
I went on vacation for 2 weeks between jobs and attended the MacHack conference in Michigan. When I came back, I couldn't find Gene, but I did find an organization chart. The two projects, Blue and Pink, were now separate organizations. I searched for my own name or my manager's name. I found Gene's name as manager of the Finder and Applications group in Pink. I found my own name as the lead programmer of the Blue Finder team, and as the acting manager of the Finder and Applications group in Blue. My manager was someone I had never met, named Gifford Calenda. I was annoyed, because it looked like many of the people I had looked forward to working with were now in a separate Pink team. Also, a bunch of people who had worked on Macintosh Plus and Macintosh II were going to join Steve Capps on the Newton project. And there was a secret project called Jaguar or sometimes Red where they were going to make a great new Macintosh using a RISC processor. It seemed like the small-team spirit of the original Macintosh development team was available if you went to work on Newton or Jaguar.
That Pink team eventually became the company Taligent. The Blue team went on to develop the next version of system software for Macintosh, called System 7. The Newton team eventually shipped a product they called a communicator. More about that later.
In the system software group, I got to do a lot of different things. I was the project manager for the Finder project. I led Finder design meetings. I wrote code for the Finder and did all the miscellaneous work on things like desk accessories and control panels so that the other members of my team could concentrate on the big stuff like the Finder and Installer.
I worked on the system patches along with others, including a great programmer named Bruce Leak. Bruce was working on something we eventually called 32-Bit QuickDraw, and we spent a lot of time together fixing bugs for Macintosh system releases 6.0.3, 6.0.4, and 6.0.5 which went along with the new computers Apple was releasing: the Macintosh IIx, IIcx, SE/30 (which should have been called the Macintosh SEx), IIci, and Iifx.
After a year as manager of the Finder and Applications group (they dropped the "acting" after a month or two), I talked with the three people managing the Apple system software. It was clear to them and to me that our large project was missing a central technical leader. They convinced me to take this type of job, but to avoid offending other people on the overall team, I chose the title "scapegoat" instead of "technical lead." It sounds silly to me today, but it seemed like a good idea at the time. My first assignment as technical lead was to work on virtual memory software for the Macintosh.
One of the engineers from the Pink group, named Phil Goldman, and his partner, Rick Daley, decided that virtual memory for the Macintosh would be easy. Since no one in system software was willing to work on it, they decided to do it in their spare time. They had a lot of spare time, since their part of the Pink project was the only one that was ahead of schedule -- they were waiting for other people in Pink to catch up. They got virtual memory to the point where it worked -- not ready to ship, but clearly demonstrating that it could be done. I took over virtual memory along with the help of someone named Joe Buczek. I worked on it nonstop for weeks and got it much closer to shippable, and then handed it off to Joe.
The Blue Meanies
Around this time, people decided that they liked the kind of general expert role I filled. The project manager for the system disks asked if we could have someone like that to help with each release. I hired a couple of people to help with this sort of thing -- the best Macintosh experts I knew. One of the first people to join the group, Chris Derossi, came up with the name for it: the Blue Meanies.
There were 2 more long years working on System 7 after I became scapegoat and formed the Blue Meanies. System 7 had this problem: instead of having a central design point like the improved Finder the project was simply the next thing from each subgroup in system software. Thus, there were well-thought-out features, like the new find added to the Finder and file system, and much-less-important, less-thought-out features, like a new way of booting SCSI devices. The project as originally conceived included much of what eventually became QuickDraw GX. I have a lot more good System 7 stories, and I'd like to write about some of those some day.
1991-1997: General MagicBefore System 7 was finished, I went to visit a guy named Marc Porat. Larry Tesler, one of Apple's chief scientists, had given him my name. I didn't know Larry well, but had worked with him just enough to know that he was brilliant; I guess he knew me just well enough to consider mentioning me to Marc. Marc was starting a project that he called Pocket Crystal. He showed me a model of a small device with a telephone and video camera on it. It had a little LCD screen. He told me about his vision for something you carry with you all the time, and use to keep in touch with the world. It seemed interesting, but a little gimmicky, and Marc's polished sales pitch also scared me a bit, reminding me of bad old days at TMQ where we had to do some fast talking and a bit of lying to get and keep business. Marc showed me a letter from Bill Atkinson that said Bill was going to work on the project to help convince me to join in. I told Marc that I was 100% committed to finishing the System 7 project, and he promised to get in touch with me when I was done.
Months later, Marc called to let me know that his project was "spinning out" from Apple, and to give me a chance to be part of it. In the interim I had heard all sorts of rumors about Apple management wanting to merge the Pocket Crystal project with Newton. Larry Tesler ended up joining the Newton project, and Bill Atkinson and Andy Hertzfeld ended up starting this new spin-out company with Marc, which they ended up calling General Magic.
Once System 7 was done, we knew two major things we had to do in Apple system software. First, we needed to beef up the operating system underpinnings of Macintosh software. Second, we needed to move from the 68000 family to a new processor with better price/performance characteristics. The big boss, Roger Heinen, in charge of the newly-formed software division of Apple told me that I could really "earn my stripes" if I worked as company-wide technical lead for this new transition to RISC. This rubbed me the wrong way. First, I thought I had earned my stripes already as technical leader of the System 7 project, the biggest software project Apple had ever done. I joined it at a time when it was in trouble and stayed with it through the time it shipped. Second, I was under pressure to give up being manager of the Blue Meanies to take on this new role, and I wasn't sure that I wanted to give up being a manager. They convinced me to give up my management duties anyway, and I got a new title, "Chief Architect of Macintosh System Software."
Since System 7 was done, Marc called me. Actually, he did something better. He got Phil Goldman, who I knew from Apple, to call me first. I went to visit General Magic. I was still naive enough to think of my visit as a social call instead of an interview. General Magic was pretty small, about twenty employees, and they had a nice building. On that visit, I met Andy Hertzfeld. I knew Andy by his reputation. A legendary hacker, he had left all sorts of things in the system software that we had a hard time maintaining. Once I met him, though, I immediately saw something else. Andy was smart and interesting. There was a lot I could learn from this guy. He had values similar to my own. After an hour talking to him, I was pretty sure I had to work for General Magic. Later, I met Bill Atkinson. It was clear that he was as brilliant as his reputation. He had a problem listening, though. The whole time I talked to him, I couldn't see any evidence that he was paying attention to what I was saying!
Unsure about the risks of working with Andy and Bill, who were both famous and infamous in my circle of professional friends, I called Steve Capps to ask his advice. I didn't know Capps well, but I trusted his opinion about Andy and Bill. At this point, there was a lot of rivalry between the Newton team and General Magic. Newton's original electronic paper vision had been refined until the product was quite similar to Marc Porat's Pocket Crystal concept. Capps himself had started a revolution within the team by making an entirely new Newton that he called the Newt that was shaped like a small notebook instead of a large piece of paper. Now both the Newton team and General Magic were making similar devices, although Magic talked a lot more about communication. Capps told me that it would be great working with Andy and Bill as long as I could hold on to my own identity and not get engulfed by Bill's powerful personality. I was impressed; he gave me good advice and didn't try to sabotage his friends by telling me not to work with them.
Marc Porat offered me the job of Director of Device Software Development at General Magic. He said that I might be overall director of all software development if that made sense later. Besides the device software there was another part of the engineering team working on a technology we called Telescript and a hardware development group. We negotiated salary and I was ready to accept the offer. Unfortunately, even though Bill, Andy, and Marc agreed that I could have that job, they forgot to consult the other members of the software development team -- also, I hadn't even met those people. The others said that they didn't need a manager. Bill and Andy were managing the team, but the way they talked about it was, "there is no manager." I went to a painful meeting where the team told me they didn't want a manager, but I was welcome to come work as a member of the team. This was the first time I had met these people. That sounded fine to me, until I went to meet with Marc. He told me that I had to take a lower salary if I wasn't going to be a director. I said goodbye and resigned myself to working at Apple for a while longer.
That weekend, Marc called me back. Andy and Bill had told him that he should still hire me with the same salary he had originally offered. I took the job, although I didn't feel quite as good about it as I did when I was going to get the bigger job and title.
My bad attitude about Magic Cap and Newton
Working at Magic was great, especially the first year. We were creating and refining both the user interface and all the levels of software for a product we called a "personal communicator." Apple heated up public interest in the category by showing a model of the Newton, describing it as a "personal digital assistant." Everyone in our team was disappointed as Apple's marketing converged on ours, though. Soon, they were talking about communication, even though there was no modem or any way to communicate built into the Newton hardware. They even named the Newton product the "MessagePad," implying communication stuff that wasn't there. We felt that they had co-opted the positioning for our product. Worse, it seemed that the Newton marketing team did this without the engineering team changing the product to match the message. We were worried about the kind of reputation personal communicators would get if the first one out didn't have any built-in communication.
A book about the Newton devlopment was published with the close cooperation of the Newton development team. The book was called "Defying Gravity". Andy Hertzfeld got a copy and was shocked by what he read in it about General Magic. According to the book, General Magic was formed to work on Telescript. Only later did we decide to make software for communicators, software that resembled the Newton software. This was wrong, and seemed to us to be a blatant lie. Marc Porat's Pocket Crystal project at Apple was all about portable communicators. The Telescript idea didn't even show up until months after General Magic had been formed. Andy took a few small pieces of paper and attached them to the cover of his copy to make the title read "Denying Reality" instead.
Eventually, I did become the director of the software development group. It was a pretty tough job. Our project was behind schedule, and some of the team members weren't getting their work done. The way Andy and Bill had structured things, each individual had a lot of responsibility and freedom. That was a good way to run things, but it made us especially vulnerable to someone who wasn't doing good work. I ended up firing two people, after working a lot with each of them to help them do their part of the General Magic device software.
Working with the software team was great. Andy and Bill were great to work with. Some of the best people I had worked with at Apple ended up working on the team: Phil Goldman, Chris Derossi, John Sullivan, Bruce Leak, Greg Marriott, and others. Also, we did some great user interface design work. Like the design of the original Macintosh, it was controversial, especially among people who had only seen the surface appearance and never used the software. To get a good understanding of the how we arrived at the design, you might want to look at Kevin Lynch's excellent materials about how we did the development. And perhaps some day I can get him to post some of that stuff so I can link to it here.
We had originally thought our project would ship in 1992, but we finished it in 1994 instead. We named the software Magic Cap 1.0. Sony built a communicator with our software; they called it the Magic Link or PIC-1000. Motorola built a radio communicator with our software that; they called it the Envoy.
The next version of Magic Cap software, Magic Cap 1.5, was finished in 1995. Sony built a communicator with it; they called it the Magic Link PIC-2000. Motorola planned to upgrade their Envoy product to this new version.
As of early 1996, I was working on another version of the Magic Cap software. Waldemar Horwat was working on it too. I'll write about that here some day and write an ending to this part of the tale.