Commodore Legends: Dave Haynie – Part I
When we talk about the C128 and Amiga, one name pops up every time: that of Commodore whiz Dave Haynie.
It is thus only fitting that after last month’s interview with Michael Tomczyk, giving us some great insights on marketing at Commodore and the success of the VIC 20 system, we now invite Dave to share some thoughts on Commodore, computing and his amazing experiences with the Amiga.
The interview will be published in 2 installments just like last month, with the first one zooming in on Dave’s career at Commodore and his work on the various systems. Part 2, next Friday, looks at “life after Commodore” and looks back with some personal views on how Commodore has influenced him.
Robby: How did you start your career in the IT industry?
Dave: I was always interested in making things. As a little kid, I used to build tree houses, forts, various downhill vehicles. I made a working radio and a magnetic induction transmitter/receiver for projects in Junior High, I taught myself programming when I was 12, thanks to a desktop HP calculator and my Dad’s login on a Bell Labs mainframe over the weekends. I took electronics each year in High School, and actually taught most of the first semester computer programming course my senior year, as our computer teacher had broken a collar bone and was out most of the semester.
So, not surprisingly, I went to college (Carnegie-Mellon) for Electrical Engineering. I added Math/CS as a dual degree, and almost had the scheduling down to toss Psychology in there too… but not quite. So I “only” took five courses in Cognitive Psychology.
During college, I worked two summers at Bell Labs. My Dad managed to get me an interview for this with an on-campus recruiter, so I was hired. It was my one real positive “big company” experience, largely because Bell Labs in those days ran as a network of departments, each one working like a small company. From College, I started out of school at General Electric in Philadelphia. I almost certainly would have gone to Bell Labs, but 1983 was the year of the AT&T breakup, and they had a company-wide hiring freeze.
GE was everything bad you can imagine about a big company, at least their branch in Philly. I was ready to leave after two months, and did, after four months. They had something like 50 people on a single VAX computer, many running simulations. After I had figured out how to schedule my stuff with higher priority than most others (spawn it as a background job, rather than submitting as a batch job… this was under VAX/VMS), I was still getting hours between results. So I also wrote a LISP interpreter to pass the time.
I finally decided to leave GE, both because it was boring and because of moral reasons. I didn’t have my security clearance yet, and they brought you in promising all kinds of things being done on the space shuttle program. But it was pretty clear that the stuff I was working on (hardware simulator) was mostly going to be used to build nukes. I didn’t want any part of that. So I sent out my resume to the one headhunter advertising in the Philly “Inquirer” who didn’t say “2 years or more experience”, and in less than a week, I had an offer from Commodore.
Robby: You started out as an engineer in Bil Herd’s team, and then rose to the role of chief engineer when Bil left Commodore. What was the life of a chief engineer at Commodore?
Dave: After Bil left, Commodore was on rough times, so it was pretty weird. They had bought the Amiga, but apparently overextended the company in the process. We had several rounds of layoffs. The first actually got rid of some “dead wood” types, “human NOPs” as the software guys called them. But it got pretty thin, and there wasn’t much direction.
I saw it was going to be pretty important to help find direction. You all know Commodore Marketing from the outside, but it wasn’t much better on the inside. In a traditional company (or so I’ve heard), you have people somewhere who think up new project ideas. That might even be their own or primary job, but at Commodore, most ideas came from Engineering. Which is cool in a way, but on the other hand, you’d kind of hope that someone in the company would be able to sell the project.
Now, I had a small problem. I was left as the senior 8-bit systems guy, but I was also using an Amiga by then. And so it was a little hard to convince everyone else that we needed to follow-up after the C128/C128D, but I didn’t have much of another course. Frank Palaia, the Z-80 guy on the C128, was also still around. So we did a “dog and pony show” version of a C256… I got it talking to 256K of memory, and Frank got the Z80 to run on every clock cycle, for a full 4MHz, rather than the half-cycle thing it did on the C128. This was well accepted, but didn’t get any green lights. I also proposed a scaled-down C128, with just the 80-column display, make it much cheaper, etc. That didn’t go anywhere either.
Chicago CES 1984 for the PLUS/4
(Picture courtesy of Dave Haynie)
Robby: One of your first tasks was the completion of the TED/264-line. Early reviews of the systems were good but it was Commodore vs. Commodore with the C64 still going very strong. How did you see this line of computers? Was it bad timing? What’s your view on the fact that they never caught on.
Dave: In a big way, I’ll claim “industrial sabotage”. Lemme ‘splain. Jack Tramiel was really frightened (well, I’m not sure the man ever actually feared a thing… maybe “concerned”) about the cheap Sinclair computers, those things that Timex was marking in the USA for about $100. The original plan of the TED project was to deliver something that sat below the C64, but was a much better computer than the VIC-20. We actually had a prototype around called the C116, which had 16K and actually a rubber-membrane “chicklet” keyboard, which was just that computer.
But somewhere along the way, the project got exploded all over the place. We had the C264, which became the PLUS/4, and the C364, which had a larger keyboard and a built-in version of Commodore’s “Tragic Voice” speech synthesizer, as we called it. At Commodore Japan, they also put together the C232, which was a C264 without only 32K, and without the user port. And of course, we did actually also ship the C16… the closest we got to the original plan in production. At one point, they had planned to offer different versions of the C232, some with maybe Logo along with BASIC, some with a good word processor. But before you know it, this “4+1″ software package appeared out of nowhere, and it was pretty heinous.
The apparent purpose behind all of this was Jack’s leaving the company. Jack left, but several of his sons were still running things here and there… and of course, the “going big” stuff started under Jack. I never got the whole story, but it’s pretty clear that this DID throw a monkey wrench into Commodore’s product line. It was, however, a good learning experience for us… many of the good ideas in the C128 were prototypes on the TED systems.
Robby: Then indeed came the C128, the “3 computers in one”. How did you see the market and the needs for the CP/M?
Dave: The point of CP/M was simple: Commodore wanted to market the C128 against the Apple II and the IBM PC. Both of those systems had available 80-column software for business that ran from a fairly conventional disk operating system. So did CP/M. And it was a fairly low-cost thing to add. That was the point.
Greg Berlin and Dave Haynie at the Tscuba Expo, Japan, 1985 for C128 production work.
(Picture courtesy of Dave Haynie)
Robby: What is in your view the key feature of the C128 that made it stand out from the rest? How would you have pitched the C128 if you had to sell it to a C64 user?
Dave: For Commodore 64 users, the main point was that here was a new “Commodore” style computer… everything you liked about Commodore machines since the PET days, very classic design with all kinds of stuff to mess around with. It has an 80 column mode with its own memory, up to twice as fast as the C64 and, oh-by-the-way, it can run all your C64 programs and peripherals (ok, 98%). At the time, the C64 was the single most successful computer model in history. Still is, last I checked (well, unless you start counting iPhones or iPads, I think). So a compatible upgrade wasn’t a bad idea, for those generally happy with the C64. And people got that — in fact, sales of the C64 bounced up once the C128 was out, just based on the idea that the C64 now had an upgrade path of some kind.
Robby: The C128 evolved into the C128D with the built-in 1571 drive and then into the C128DCR. It sure was a bit of a shock when the reports came in that it was showing compatibility issues. What was the driver for having the “cost reduced” version out and were the compatibility issues to be expected?
Dave: The driver for any cost reduced product is one thing: cost reduction. There were five major redesigns of the C64 over the years. Why? The “E” board model, in a case, in a box, with power supply supposedly cost Commodore $35 to make. When you sell a million+ units a year, even saving a dollar is money in the bank. And keep in mind, in those days, you didn’t exactly replace your computer models two or three times a year as they do now.
I really didn’t have much involvement in the C128D-CR… most of that effort was merging the C128D with the 1571 disc drive into a single motherboard. The original C128D used the same motherboard we had in the C128, and a separate 1571 board. So apparently, in merging the two, some compatibility issue was affected.
Keep in mind, that was a tricky thing in those days. You think the C64 ROMs were software, but they were really an effective part of the hardware description of the computer. That’s why the C128 banked in a real C64 ROM when running in C64 mode. And we even changed the character set. I was the main guy responsible for compatibility… I tested 3rd party stuff extensively, and even when they did things that we didn’t like, we changed the hardware where possible to support not just what should be supported, but what could be.
So the software guys had actually cleaned up the C64’s character font a bit. But we found even that could break things. One important painting program, for example: a graphics program from Island Graphics. The start-up screen of that program painted a large “Island Graphics…” and some other stuff on-screen. Apparently, it read that font from the C64 ROM, and when the C128 characters were in place, they went to dot the “i” in “Graphics” and missed… causing the whole screen to be flood-filled. And that set off a chain reaction of failed graphics commands that meant, while the program worked, it took over half an hour to start. So we put in a double-sized character ROM… C64 font in C64 mode, C128 font in C128 mode.
Robby: Moving on to become the primary engineer on the Amiga, what was your elevator pitch on the Amiga?
Dave: World’s most advanced personal computer. Back then, it was true beyond question. And that’s not say everything we had in there was perfect, but when you added it all up, the others were found wanting. And I did not become THE primary engineer on the Amiga.
Basically, after those C128 follow-ones failed to catch attention in West Chester, another project did: the Amiga 500. George Robbins, Bob Welland, and several chip designers had conspired to lower the cost of the Amiga chipset, and put that into a traditional Commodore “flat” case. Remember the aftermath of the Amiga buyout, the layoffs, etc. Another casualty was the Commodore 900. There had been several attempts to build a high performance 16-bit computer, running a UNIX-like OS, with high resolution monochrome display. That was the Commodore 900 — Commodore’s answer to the Sun 2. Bob and George were actually the third team to work on it, but the first to actually get it working. So naturally, the C900 was cancelled. And they went to work on finding something to keep them employed. The Amiga being the most interesting thing, that’s where they went.
Anyway, I came onto the A500 project in the summer of ’86. The plan was for me to get into the design, learn it, and eventually take over, so the high-end guys, George and Bob, could go do some high-end stuff. Only thing was, George really liked the A500, it was his creation. So when we got the job of doing a cost reduction of the German Amiga 2000 design (in short, adapt the A500 chip architecture, Fat Agnus, and integrate the rest of the PALs from the Zorro expansion bus), I wound up being the guy. And so I kept on, usually, as the main or one of the main engineers on high-end Amigas ever after.
Joel Tessler and Dave Haynie in Basel, Switzerland, 1990 – Amiga show
(Picture courtesy of Dave Haynie)
Robby: How did you see the competition from Atari (i.e. the STs), Apple and the up and coming DOS market
Dave: Yawn, mostly.
MS-DOS was doing well, but because of IBM, not any particular merit of the products. And of course, because by then you could clone the IBM PC. So dozens of companies who could never have come up with a computer from scratch like we did, could now enter the personal computer business. Naturally, that made them a potential threat. But they had not yet started to integrate in silicon, the era of the cheap PC was still years away. The best things from the PC: simple expansion bus that anyone could design to, and that $150 add-on FPU. That FPU made spreadsheets run 100x faster, and that alone justified that PC in business. It meant that, rather than just do your books in a spreadsheet, you could actually model complex things. I was very interested in getting an FPU in the Amiga. And of course, Zorro II was pretty easy to design for, and better still, easy for customers to use.
Apple was interesting. The MacOS, deep down, was a horrible, primitive thing. But the stuff on top was pure sex. They had (and still have) an amazing attention to detail that practically no one else bothers with. They had some very slick looking stuff, but the thing I found most dangerous to us was the fact they had retargetable graphics. Pretty early on, Commodore was backing a graphics card project at the University of Lowell, MA. This eventually saw a little production run as the Amiga 2410. The problem was, AmigaOS couldn’t use it… only Amiga UNIX. I saw this dependency on Amiga chips as a serious problem.
There was really nothing interesting about the Atari ST, other than its history. I did mention the C900 team. At one point while at Commodore, Jack had a few of the then C900 engineers (I guess they were the second team on the project) meeting “off-site” on some super secret project. This was probably late 1983 or early 1984. Shortly after Jack left, those guys all followed, and of course surfaced at Atari. With a new 68K computer that looked an awful lot like the second spin of the C900. Well, other than the fact it actually worked this time :-)
Robby: What’s in your view, the most groundbreaking Amiga innovation?
Dave: There was many…. that’s the main reason the Amiga was so amazing when it was first released. Most new computers might at best do one thing better. The Mac was seen as amazing due to the megapixel GUI display, most ignored the fact it was almost too slow to actually use, and that 512 pixel display didn’t deliver an honest 80 columns. The IBM PC… well, it was from a “real” computer company, that’s about it.
But the OS was the big, big win. Personal computers had been kind of re-inventing the 1970s, like IBM and Microsoft essentially cloning CP/M to make PC-DOS, in the early 80s. Or Apple building a GUI over a glorified Apple II DOS. But the AmigaOS was amazing at nearly every level: great architecture, very efficient multitasking, a powerful, device-independent DOS, command shell and GUI, automatic device configuration (which is still, IMHO, the correct way to do this kind of thing… Windows 7 still does it wrong).
This concludes part I of the interview. Next Friday, Dave looks at “life after Commodore” and his experiences at Nomadio and Fortele and gives us an insider’s account of the last days of Commodore. All this, and the evolution of the AmigaOS after Commodore, in next Friday’s part II of Commodore Legends: Dave Haynie.