146
edits
Changes
Jump to navigation
Jump to search
Created page with "This is a transcript of an audio interview. This transcript may contain errors - if you're using this material for research, etc. please verify with the original recorded inte..."
This is a transcript of an audio interview. This transcript may contain errors - if you're using this material for research, etc. please verify with the original recorded interview.
Source: ANTIC: The Atari 8-Bit Podcast
Source URL: http://ataripodcast.libsyn.com/antic-interview-44-joe-decuir-atari-400800-designer
Interviewer: Randy Kindig
RK: Hello, and welcome to another interview edition of Antic, the Atari 8-bit Podcast. I am Randy Kindig. I am extremely honored to bring to you today, an interview with one of the true pioneers of the personal computer, one of the primary designers of the Atari 400 and 800, Mr. Joe Decuir. Joe worked closely with Jay Miner, Steve Mayer, and others, to design a computer in the days when personal computers were just in their infancy. Joe shares with us the thoughts and reasoning that went into the design of the 400 and 800, and thus brings all of us a little closer to being part of an amazing time in computer history. Please enjoy.
RK: Okay. I’m here with Joe Decuir, one of the designers of the Atari personal computer system. Joe, how are you today?
JD: I’m having a good time.
RK: Good. Can you tell us a little bit about where you currently live, and what you currently do today?
JD: I live in Issaquah, Washington. I moved from the San Francisco bay area to the Seattle area, to work for Microsoft in 1992, because the company I worked for went out of business, and I needed a job, and I dragged my wife and kids up here, and I’ve been here ever since. What I do for a living, my “day job” is, I’m what they call a standards-architect, at a company called CSR, as in Cambridge Silicon Radio. What I do for a living, is I write widely-distributed public engineering standards. I’m sure some of you’ve heard of Bluetooth, and you’ve certainly heard of USB, and I’ve worked on a whole lot of other stuff in the last 28 years. Made a difference in the world. That was fun. My main avocation is this global humanitarian technology conference that I’m going to be running in Seattle in October.
RK: What’s that about?
JD: Well, I work for… I’m a volunteer at an organization called IEEE, that used to stand for Institute of Electrical and Electronic Engineers. But it’s decided to rebrand itself as Advancing Technology for Humanity. We thought “hey there’s a lot of people out there that don’t have first world problems.” At least 4 billion people in the world live on a dollar and half a day or less. This conference is about them. It’s about rural power; it’s about rural healthcare; it’s about rural communications; it’s about disaster response, and that kind of thing. I really believe in this stuff. I had to choose between doing a good job at my day job, and doing a good job with this, I would do this. This makes more difference to more billions of people.
RK: Yeah, that definitely sounds pretty neat, I’m sure that that’s a lot of fun to work on.
JD: Yep.
RK: All right, well, I know you’re working on a book, let’s If we could, jump in to the Atari here. I know you’re working on a book, having to do with the Atari personal computer system. Could you maybe tell us a little bit about that, what that’s about, what the purpose is, that sort of thing?
JD: Yes. As a part of this I have a lot of books and other documents in my garage, and my wife would like me to get rid of them, and to do that, I first-of-all want to turn them into a book. The MIT Press, a few years back, did a terrific book on the Atari 2600, called “Racing the Beam.” It explains pretty well, how we designed it, and a lot of the cartridges that went into it. Separately a few years later, somebody wrote another great platform book from MIT Press called “The Future Was Here” about the Amiga. The book I want to write is in between. It is a sequel to “Racing the Beam,” and it is a prequel to “The Future Was Here.” The first chapter in this book will talk about what lessons we the engineers learned from the successes and failures of the 2600, and how we chose to address those in our second system. At the end of the whole process, the last chapter in the new book will be a prequel, what lessons we learned from the Atari personal computer system, and how we address these in the next system. So those starting and ending chapters will tie the three books together.
RK: So what’s your time schedule for getting that out the door, do you have one right now? Or are you just trying to work on it as you can?
JD: I have 45,000 words in the document so far. Maybe it’s 45,000 characters, whatever, it’s large, but it’s not done. There are two things I want to do, and they’re distinct. One of them is I want to find more of the original participants, and tell the story in their words, rather than just mine. The most important is a fellow named Doug Neubauer, and I really want Doug because he is still alive, unlike some of the other designers, he was the main designer of the POKEY chip, which is all the non-video I/O, and he also wrote the definitive game cartridge for this system, which was Star Raiders. If I can find him I’m great. He lives in Oregon apparently, but he hasn’t returned my phone calls. The other thing I want to do, in this book, is this machine was designed explicitly so that all kinds of people could port games and create games on it, for it. The Atari 2600 was not a development system, it was a play system. The Atari 800 system was designed so that you could write games for it, on it, and then sell them. A big chunk of the middle of this book needs to be about how you would actually take handful of popular coin-op arcade games and port them to this platform. I actually want to show the design at some high level, and some of the code. So that people can go off and write their own.
RK: Yeah that sounds very interesting, that certainly would be very different than you typically see in these types of books. So if you don’t mind, can we talk a little bit about the history, I know that you were intimately involved in the creation of the Atari personal computer line. Can we talk a little bit about what was the impetus for creating that, and why did Atari decide to go forward with creating a computer line?
JD: Well the first thing you have to understand, is we started off in the games business. We started up in the coin-op games, and then we started bringing things home. We brought Atari Pong home first in 1975. That said, we then wanted to then bring the complicated arcade games home as well, and the only question is how are we gonna do it. We designed a special and complicated system that was a line-buffer only, not a frame, and that made it achievable, financially, in 1977, when the product came out. But we knew, when we started the process, that semiconductor technology was going to catch up with us, and it was gonna become practical to design systems with frame-buffers rather than line-buffers, and a lot more horsepower, so it would be possible to bring home the things that we were pioneering in coin-op. We thought, in fact, when we started the 2600 process, we thought it would sell maybe for three years, you know like ‘79… no ’77, ’78, ’79, maybe ‘80. We knew… we thought that we had to replace the product, before our competitors did, we assumed that they would, and they certainly tried. So stepping back, we wanted to build the next great game system, and we wanted to build a computer. There was a big ongoing question about if they would be the same project, or different projects. We thought, erroneously as it turns out, that the great home computers had to play great games. The market didn’t support that. If you look at what happened in the 1980s, the Macintosh’s and the IBM PC’s and their clones, took over the personal productivity world, and the Nintendo NES’s and its successors, took over the home console business, and so they were considered separate things. In the late 1970s, we didn’t know that yet. So we tried, twice, to build a great home computer, that would also play great games. The second instance was the Amiga. Which was a monster game console, disguised as a personal computer. In retrospect I wish we had sold the Amiga as a game console, because it would have blown the doors off of the NES. The NES caught up with it with the super NES in 1992, is when they finally caught up to the Amiga and what they could do. But I’m digressing.
RK: So, yeah that’s fine. So as far as the home computer competition, though, at the time you guys decided to build a home computer, what was your biggest competition, and what did you think you could do better than the other computers that were out at the time?
JD: Those are two separate questions, and I’ll try to cover them both. First of all, I was one of many people who started working on S100 machines, in 1975 and 1976, and I was in the homebrew computer club and I hung out with those people. The first machines were the S100 machines, and then Wozniak introduced the Apple I, let’s see at April 1976, at Stanford. We thought “there are going to be personal computers, the question is what are they going to do.” The second wave of personal computers was appliances, the Apple II was the best example, the Commodore PET was less so, the TRS-80 wasn’t bad for what it did. All three were pointed at productivity applications. An old friend of mine, Bob Frankston, wrote VisiCalc for the Apple II, which was the killer application at the time. So we were thinking “we were heading in that space.” A problem that we knew were going to have, and we didn’t know the FCC was going to change the rules, we were designing for home computers… home users, not office people. So we figured that the home users weren’t going to buy separate monitors, ‘cause good color monitors didn’t exist in the late 1970s. They were precipitated to keep up with the IBM PC which came out in 1982. Right, so we’re five years before that, we don’t know this is going to happen. We designed therefore for home televisions. The 2600 was designed that way, and the 800 system was designed that way. We said, “given that we want to do personal productivity, how many characters can we put on a screen?” The physics, if you drive a home TV, you’re stuck with the upper limit of the color clock, which is 3.58 MHz. So that’s your top clock rate. So at best you can do 40 characters across, or 160 pixels across. That’s the best you can do. Does that make sense? We were driving at that. So the question is “what are we going to do that they can’t do?” Well, we’re going to be driving home TVs, which the Apple was not designed for. In fact, if you want to an Apple Store, and bought an RF adapter from a separate company, you could drive a home TV, and the FCC didn’t go after Apple, because the FCC hadn’t noticed yet. But we were Atari, and we were stuck with having to connect to TV leads, which were also indirectly connected to the antennas, that could cause interference with your neighbors, we didn’t want to do that ,the FCC would crucify us. So we designed systems that lived inside of what I would call a Faraday cage, are you familiar Faraday? Famous scientist from 200 years ago. If you basically take the circuit board, and wrap it in aluminum casting, then very little radiation gets out of that box, that’s the way their designed. Which then stuck us in a difficult problem, because we didn’t have slots, like S100 boxes or Apple II’s, that means we had to do external peripheral expansion. That means we had to make up a serial bus, the first one that I was aware of, and connect it to separate external peripherals which had their own processors in them, plus a printer, or plus I/O, or plus a disk drive, or whatever. So we were breaking ground, there. It turned out, in retrospect, the FCC didn’t drop the ball on Apple for another several years. That means that we had a price disadvantage selling as a full computer against the Apple. By the time the FCC did change the rules, we then had to face Commodore.
RK: So, can you help answer, once-and-for-all, where the names Colleen and Candy came from?
JD: Ah! They were Atari employees.
RK: They actually were? Ok.
JD: They’re now called an administrative assistant, but at that time they were secretaries, but that’s what they did.
RK: Were they picked for some particular reason or?
JD: No. Marketing told us their names, but they were employees, I’ve actually seen the badge of Colleen, she was a redhead. I think they both worked in the coin-op part of the company. They’ve, you know… women from that time tended to change their names when they got married, so it would be rather difficult to track them down.
RK: So, if you don’t mind can we talk about some the critical design decisions that were made along the way, and why those were made?
JD: That’s what I’m here for.
RK: For instance the processor. Why did you decide to go with the 6502?
JD: Ok. You probably are not a scholar of processors from that period. But until you jumped to 64-bit stuff, I think the 6502 was the best 8-bit processor at the time. It is architecturally superior to the 8080. It was architecturally superior to the 6800. I think it was still better than the Z80. It was pipelined. It was little-Endian. Which means if you’re doing… You can in one instruction… Let’s suppose you have an instruction that’s two bytes long, the first is the opcode, the second is the offset in zero-page. That instruction can point to a 16-bit number in zero-page, which is either pre-indexed by the X-register, or post indexed by the Y-register, so you can do address generation and calculation on the fly. The 8080 can’t do that, the 6800 can’t do that. Because of those architectural features, it was possible, and let me underscore, capital-P “Possible”, to build the 2600 with a line-buffer only, and not frame-buffer. Given that, we had a lot of them in-hand, they tended to run fast. We still liked them architecturally better than we liked the other processors, so we used a faster version in the computer, but we still used basically the same machines. We tried and designed a… what I would call a 6809. We did a complete design for it, we scouted it out with some of our silicon vendors, and they said “well it was extra risk, to you know, administrative risk in the design of the product, to go for a 6809 with 16-bit instructions.” So we killed that project. Later, basically the same design became what is called the 65816, which was used in the Apple IIc and IIe.
RK: Ok. The video. Now I always thought the Atari computer line was, you know, really excellent on a TV set, you could tell it was optimized for the TV.
JD: Yes it was.
RK: You talked about this a little bit already, I understand. Can you talk a little bit more about some of the design decisions that were made there?
JD: Well first of all, again, we’re split between trying to build the best game player we can build, and a home productivity thing. So as a home productivity thing, it needed have enough memory to run the code itself, and then have buffers to build the things that you were working on. For that we were basically thinking monochrome characters. So we came up with a monochrome character set that would do 40 across that would work. But having designed that, we had some other things that would be useful for other stuff. So thinking about as a game machine, as we built more and more complex VCS cartridges, we kept wishing we could do more and more complicated backgrounds. So we basically took monochrome characters 40 across, and took the same hardware, and made character… you know multicolored character maps, where there are four pixels across, and eight pixels tall, and they’re four colors. So we made up color-character mapping. Having done that, we can now build very complicated backgrounds with very small numbers of bits, because a lot of bits are being reused on the fly. Now having done that, we thought “hey”, let’s build a… we noticed that the successful VCS game designers, their processors were busy following the beam down the screen, so we decided to design a simple graphics processing unit, as part of the system, that would follow things down the screen. So instead of having a fixed-resolution bitmap, or character-map. We made it so that it changes modes, on a line or a group-basis. It could have a few lines of characters, and a few lines of bitmap, and a few more lines of characters, it could do all kinds of things, and change them on the fly. We then went further and made it possible to make them scroll horizontally and vertically. If you’ve seen the 2600 game “Adventure”, you know it moves, you know, from one frame… you go off the side, and then it jumps to the next frame. In the new machine you can smoothly scroll arbitrarily large images. It turns out, by the way, that Nintendo had to pay royalties to Atari later, ‘cause they replicated that feature in the NES.
RK: Oh really? I didn’t know that.
JD: Really, it’s not widely known; and I wouldn’t wave that in front of a Nintendo lawyer.
RK: I don’t imagine they care too much today.
JD: But it’s true. So yeah, we were trying, like I say, trying to build a machine which simultaneously can do personal productivity, and really great gameplay. There aren’t many Nintendo games that couldn’t run on a 5200, except we didn’t designed that way, and we didn’t package it that way.
RK: What about sound?
JD: Sound. Sound was a case where we didn’t do very much new design. We had two channels of sound in the 2600. We went four channels, with more frequency control, in the 800/400 system, but we basically took the old design and doubled its’ size. We didn’t put a lot of effort into sound there. We went much farther in the Amiga, but that’s a another story for another time.
RK: What about the I/O bus and the peripherals?
JD: Now that was the case where, again, we are forced by federal regulations, to minimize radio=frequency interference, so we couldn’t build… If you take an Apple II today, in its’ plastic case, it makes too much radiation; you can measure it, it’s now illegal to run an old Apple II. Just that there aren’t very many of them, so they don’t make a lot of radiation, right? So we at Atari were forced to have much lower RF output. It turns out, by-the-way, that ham radio people, love Atari 8-bit computers because they’re the quietest computer that you can buy to hook up to your six-band rig.
RK: Oh really? The 800 specifically because of the Faraday cage?
JD: Well yeah. It’s the RF-quietest computer that you can buy from that period. PCs make more radiation. Macs make more radiation. Notebooks are pretty quiet, again, but it took a long time to develop stuff like that. Anyway backing up: so we are thinking we don’t get to have slots, and that drove us crazy. What are we gonna do if we don’t have slots? Well we’re forced to maintain lower radiation to have from a low radiation serial bus to connect to peripherals. So we had to design a disk drive that had its’ own processor and it. In a peripheral adapter with serial and parallel ports with its’ own processor and it. And a printer that had its’ own processor in it. We were stuck. It worked, but it didn’t have the costs that an Apple II system where you plugged everything into the back bus. What we didn’t do is do a redesign, and incorporate all of those things internally, on an internal unexposed bus. Now in retrospect I wish we had built an 800-series machine with one built-in disk drive that was right on the bus. That would’ve had much lower costs than the version with the external disk drive. But by the time we got to those kinds of questions, I had left the company.
RK: Or if that was the intent of like, what was it called, the 1450 and the 1450 XLD?
JD: Yes. The 1450 was a great idea that was two years or three years too late. We should have brought it out in competition with the Commodore 64. In fact, if you look at the design of the Amiga 500, it’s the same design; except with a much more powerful system behind it.
RK: So what about the specialized processors that you decided to put in there?
JD: Ok. Well we had… we wanted to, like I say, support great video games, and personal productivity and adapt it to a TV screen. We split the graphics processing function into two pieces; the ANTIC chip, the name of this podcast, was the device that generated addresses, and figured out what to display, and that George’s TIA, was the device that captured the data, that was fetched by the ANTIC, and then shoved in to the TIA to display on the screen, so they were mated pair. The third part was POKEY, which was non-video I/O: that’s where we put the serial-bus I/O, that’s where we put the sound, that’s where we put keyboard, that’s where we put potentiometer-scanning. The Atari 800 has exposed, on those four ports on the front, it can handle basically two axis times four controllers. Very few people ever use that feature, but it works.
RK: Two versions of the computer. So at some point along the way, you decide you’re going to have two different computers can you talk about that for a little bit?
JD: We were carrying the schizophrenia of trying to do a productivity machine, and a video game machine, all through the process. We had designed a real computer with a full-stroke keyboard; that was the 800. Then we had designed this game console, which was going to be the 400, and then Doug Neubauer came along with the definitive video game which needed keyboard control, and that forced management to put in a membrane keyboard on the 400, so that Star Raiders was playable. Having done that, we ended up with basically a game console that can play complex stuff, and a real computer that can also play games, and internally from a software standpoint, they’re identical.
RK: As far as the design of the two different computers, I’ve heard from Curt Vendel and Marty Goldberg’s book, that the 800 is really designed like a futuristic typewriter, right, is that true?
JD: Well it looked kind of like a typewriter. Manual typewriters from the period had a keyboard similar. But so did, you know… A keyboard that’s not much different, ended up in the IBM PC. If you look at keyboard that was attached to the original Mac, it was a similar full-stroke keyboard. If you look at the... There was a company that did a computer called the “SOL” , that was an S100 box, it had a similar keyboard.
RK: I understand the 400 was actually, originally designed so that you could put it in your lap.
JD: Not really.
RK: Not really?
JD: I wouldn’t do that. Remember to play games on it, you were supposed to be using those controllers, ‘cause it used the same controllers as used on the Atari game consoles previously. In fact the same controllers work on the Atari’s; they work on all the Commodore… C64’s and related stuff; they work on the Amiga’s. Same controllers.
RK: As far as system software goes, what were your thoughts when you’re designing that?
JD: Well first of all I wasn’t one of the… I wasn’t the main designer of the system software. We hired a bunch more people to help do this. What we did, was first of all we knew that the machine was gonna ship with software, unlike the 2600 which has nothing in it. Have you ever seen the cartoon that has a Sega, and an NES and an Atari 2600 sitting there? And they’re talking about the good ol’ days, and the 2600 says “I have no memory.”
RK: I think I have seen that one, yes.
JD: It’s pretty funny, anyway it’s true. The thing had no memory. This machine had memory. The base machine had a generalized BIOS, and by BIOS do know what I mean by that? Basic Input Output System. It’s abstracted things so that you can connect the display to other things; to the disk, to the tape, to the serial ports, to the printers; it abstracted stuff, which made it easier to write…. you know, it was middleware. I’m assuming you know enough about the history of operating systems to understand what we did here?
RK: You know I do, I don’t want to assume that our listeners all do. So, yeah… if you don’t mind explaining it a little bit, that’s great.
JD: Well imagine that you build a basic input output system, which from the point of view of the application software, can write to the screen, or the display, or the keyboard, or a file on the disk, or a file on the tape, or a serial port, all in exactly the same structures. Because what operating systems do, is make it so that, from the point of view of the application software developer, you don’t need to know how the underlying stuff works. Right? Way back in the 1970s, and early 80s, things like MS-DOS and CPM were abstracting how printers work, so a word processing program could print to a lot of different things, and the end-user didn’t have to know which kind of printer it was driving. Operating systems have been abstracting everything ever since. The whole DirectX operating system extension in Windows is to make it easy to write video games, and the DirectX abstracted all the different kinds of display hardware and controls. As an aside the DirectX system was encapsulated into the first Xbox Microsoft game console. So that’s a story about abstraction; about making life easier for the applications developers, and a point I would make here is, you really want to please the end-users, that’s how you make business succeed. The applications software developers, in my opinion, are the closest proxy you get to the end-users. So if make life easier for them, you make lives easier for yourself as a hardware and system vendor. I work in the “Internet-of-Things” business today, and that’s the approach I’m still taking. Trying to harmonize the upper layers, and make it easier for the applications writers. That was probably a longer answer than you were expecting.
RK: That’s okay. So then, you outsourced the development of the Basic language right?
JD: Well, we had local people that we outsource it to, but we specified what it was, and they wrote it for us. The Basic language also had a 2K piece, which we included in the system, which was all of the floating-point and mathematics routines, so those could be used by a Basic cartridge, or the built-in Basic, or they could be used by another high level language like C, or any other language that you gonna write for this machine. Yeah, now you made a… had a comment about the right cartridge slot. We were making ROM-based systems, but we imagined that for the full computer, you might have a master cartridge of some kind, let’s suppose you have an educational cartridge, and here’s the master in the left slot, and then it might have a number of subsidiary cartridges that you would buy from the same company, and plug them into the right slot. In fact nobody ever bothered with that, that I’m aware of.
RK: There were few right cartridges ever created. I know we did that on a podcast once, and there were maybe a dozen, I mean not many.
JD: This is why... We were… everyone was trying to design to a market. They may have enough features; they may not have enough features. This was a case where we anticipated a need that didn’t show up.
RK: Sure.
JD: And you never know when that’s going to happen.
RK: That’s true. That’s true. So I guess that kind of summarizes a lot of your design decisions. How much did cost drive a lot of these decisions? Or was that… did that even factor into it?
JD: Our rule of thumb with the 2600, which we are replicating with the 800 system, was that the retail cost was probably three times the bill-of-materials. My experience since then, that’s probably still true. We were... Our goal was to build the best we could build, and then figure out how to get it to fit, rather than here’s our fixed-cost target were going design to fit into that. This is one of the reasons that the book is named “The Second System”, because in classic… If you have not yet ever read “The Mythical Man Month”, you should. There’s a chapter called “The Second System Effect.” To succeed, you have to keep focused, and ship on-time, and on-budget. Otherwise you fail, there a lot of different ways for a company to fail. But as you do all of this work, a frequent danger is that you’ll think “well, that’s a great idea but I’m going to save it for later”, “and I think I’ll save that one for later”, “and I think I’ll save that one for later”, so you get a chance to build your second system. The question is how disciplined are you at including the things that need to be included, and not including the things that didn’t. Now… If we had… If the FCC had changed the rules earlier, we would’ve built a machine, sort of like an overpowered Apple II. We would’ve built something with slots, which would’ve made all the peripherals a lot cheaper, and we would have been competing successfully with Apple, but they didn’t change the rules in time; we had to put a Faraday cage in there. Then when they did change the rules, in my opinion, Atari did not respond aggressively enough, to take advantage of the lower costs in packaging.
RK: So how much were you… I know you were involved in the development of the SIO bus. How much were you actually involved in the creation of the actual peripherals that ended up coming out for the Atari’s?
JD: Not very much, ‘cause I was up to my neck in trying to get the main chipset to work, and you know, doing all of the other software work I was doing. We did, early on, reach the conclusion that we needed a serial bus, and that I was one of the small number of people that wrote out the specifications for what would be on the serial bus. Then we incorporated some of that into the design of the POKEY and that guys like Scott Scheinman took the bus and went a lot further with it, and he was intimately involved in designing some of the peripherals. I wish.. I want to find Scott Scheinman as one of the people to include in the book.
RK: I was gonna say he would probably make a good interview, too.
JD: If you can find him. If you can find them I want to talk to him.
RK: Okay, then maybe we can help each other out with that, I don’t know.
JD: Yes, I want him and Doug Neubauer.
RK: What has been your involvement with the Atari history books that Kurt and Marty are working on?
JD: I’ve been corresponding with them for years. I have fed them a lot of information that’s been in the Atari Museum, and some of it’s been in there, ended up as photos and anecdotes in their book. I consider the Atari as you know “Atari: Business is Fun” book to be pretty much the definitive social history of working at Atari. Now they’ve been struggling to create the second book, which was “Atari Corp.: Business is War.” I didn’t have anything to do with Atari Corp.
RK: I understand you were a member of the homebrew computer club back in the day.
JD: Yes I was.
RK: What was that like?
JD: That was… oh God that was fun. Of course it was for hobbyists’, right? You know I started… I was working in medical research into late 1975, but I heard about the homebrew computer club, and I drove down a couple of times and started meeting these people. Through them I met the people who hired me at Atari, actually. Then I was still hanging around… I didn’t have a wife and kids yet… so I had discretionary time. So I was there in April 1976, when they introduced the Apple I. I think I probably stopped going there by 1977, but I did go to the West Coast Computer Faire were they introduced the Apple II.
RK: Oh that would’ve been interesting.
JD: It was, oh yeah.
RK: You still have any of materials from that show?
JD: I probably lost it all. This is some of the stuff my wife is trying to get me to get rid of from the garage.
RK: Oh really? Well if you ever... If you find any of that stuff, and you want to… One of the things were trying to do with this podcast as well, is to get things like that archived, so that they’re not lost to history.
JD: I have scanned some things. You can go on web search and find a scanned copy of my original 2600 programmer’s guide for example, you can get that. Like I used to have some of the mylar printouts that you use to make the chips, and I sent those to the History Museum, probably 10 years ago, so they are keeping track of some of that stuff. Now I have a few copies of Atari PCS and Atari VCS, and a handful of Amiga’s, because I even have children that I want to give them to. I’m in touch with the people at the… there’s… Paul Allen has a museum, where he has a lot of old computers that are still working. It’s called let’s see, what’s the name of it…
RK: The Living Computer Museum
JD: That’s it. I’ve been there. They’ve got great stuff. In fact they’ve got 2600s and 800s and 400’s and Amiga’s, still working.
RK: I mean… so you’ve met Steve Wozniak, and Steve jobs. I know that Steve Jobs had some involvement with Atari. Did you meet Steve as part of that?
JD: Yes he did. Several times. I had to turn him down several times for a job. The first was in February or 1976, which is when he snatched Rod Holt from my Atari team. Rod Holt did the design for the switching power supply in the Apple II. I turned him down again in 1977, and I turned him down again in 1979.
RK: I interviewed Al Alcorn, and he was going on-and-on about how Steve was luring people away from Atari.
JD: Right, and I stated it. I chose to work at Atari, rather than at a medical electronics company, in late 1975, because I thought that Atari would teach me a lot of stuff that I didn’t already know; whereas the electronics medical company would feast on what I already knew. I think you made the right decision. I got to design lots of software, lots of systems, lots of integrated circuits. It was a great time. I miss parts of it, but the company changed out from under me, as companies often do.
RK: How did you enjoy working at Atari, though, everybody we’ve talked to just really seem to love it?
JD: I used to go to work thinking “I can’t believe they’re paying me to do this.” I expected to come into work, and find a little, you know, coffee can with a coin slot in the top, that I should put quarters in to keep working. It was that much fun. I figured that out there in the world, there are probably hundreds of people, who would kill me to take my job, likely they didn’t know who I was.
RK: Yeah, well that’s great, that’s the kind of job you want, isn’t it? So you left you said in 1979?
JD: I left in the middle of 1979. Because some friends of mine and I had decided to start a communications-oriented company. I had been reading IEEE publications, and I had been reading about the French Mini-Tel system. You’ve probably never heard of this, but the French government was trying to reinvent telephones in the ‘70s, and they built a system with very cheap terminals, with character… characters and graphics characters, and simple keyboards, simple modems, and they built a computer system behind it. They were prototyping, the applications that made the World Wide Web explode in the 90s. So the French were almost 20 years ahead of their time. But they prototyped all this stuff, and I read all that, and I said “oh my God the future is in communications.“ So I left Atari to form a company involved in designing modems and other telephony-oriented peripherals, and ways to connect them to computers.
RK: So had the 800 and 400, they had probably just been announced?
JD: What they had been announced. I stuck around at Atari until production was working. When the production yields got high enough, is when I felt I was willing to leave. I wouldn’t leave until the production yields were good. So, in fact, as a side piece of the work I was doing in 1979, as I wrote a bunch of the system diagnostics that they would use in the factory to tell if they had a problem and where to fix it.
RK: Is that some of the information that ended up like in the field service manuals?
JD: Yes, although I don’t have the original copies of those anymore, I wish I did,
RK: Do you have any eight Atari 8-bit computers today?
JD: I have several. I have three original 800s. I have an 800 XL. I’ve got a 600 XL. I’ve got an Atari Corp. 65 XE. I have a bunch of peripherals for these things. I went out and bought, recently, a little adapter that takes SD cards, and emulates a disk drive. Because I want to use all those tools when I finish the book, because I have to prototype a bunch of software.
RK: Is it like the SIO2SD from Lotharek?
JD: Yes. Yes, that’s exactly what it is.
RK: Yeah, that’s a very popular one, I’ve got one of those as well. Those are really nice. It’s amazing how much you can improve those machines today, using something like that.
JD: If I had a magic wand, I would build… put together a little company that built a cost-reduced, and simplified, PCS system, set up to write new games on. But what’s happened is people have written emulators for PCs that will do this.
RK: Right, right.
JD: So I’m hoping that some of the audience for the book, will actually, with either an original machine or with an emulator, try creating the games that I’ll be describing; and then have fun in new ways that they wrote themselves.
RK: I guarantee you some people will do that.
JD: I don’t have a 400 thought.
RK: You don’t have a 400?
JD: Nope. I have… I wanted the real keyboards. I have been to a few classic games expos, and given some talks. I can send you some of the stuff that I’ve written. You saw the one that I gave a couple years ago in Portland.
RK: I did, yes, I watched that video.
JD: I’ve been invited to one that’s in October, in Portland this year, and I’ll probably go to be on a panel with Dave Crane, who is a great buddy.
RK: Oh excellent, that’s probably the Classic Gaming Expo.
JD: That’s the one, yeah. It’s on my calendar somewhere. It’s the week after this conference I’m running on humanitarian technology. For me that will be a vacation…
RK: So is there anything else I should have asked you Joe? Or that you’d like to say?
JD: You asked me a great bunch of questions, and I sort of don’t know what else to offer at this point. If I was retired, I would probably running some more games. Example: see, there’s a guy named Ed Friis. And he used to work at Microsoft, and he’s the original writer of the game Halo. He ported Halo back to the Atari 2600; I have a copy of that cartridge. I would like to see what happens to port Halo to the PCS. In fact, maybe I’ll ask Ed Friis permission, and maybe I’ll make that one in the book. He lives close to here; he’s a friend of mine.
RK: Excellent. Well, I didn’t really have any other questions, unless you had something else you wanted to say.
JD: Nope. If there’s nothing else I can do, you can correspond with me later. Thank you very much Randy.
RK: Well thank you very much for your time I really, really appreciate it. I think this is going to be very interesting for everybody, and you have been excellent.
JD: If you get more questions, you can forward them to me, and I’ll answer them as best I can. It’s a promise
RK: Ok. Very good.
JD: Cheers dude.
RK: Thanks again. Bye.
Source: ANTIC: The Atari 8-Bit Podcast
Source URL: http://ataripodcast.libsyn.com/antic-interview-44-joe-decuir-atari-400800-designer
Interviewer: Randy Kindig
RK: Hello, and welcome to another interview edition of Antic, the Atari 8-bit Podcast. I am Randy Kindig. I am extremely honored to bring to you today, an interview with one of the true pioneers of the personal computer, one of the primary designers of the Atari 400 and 800, Mr. Joe Decuir. Joe worked closely with Jay Miner, Steve Mayer, and others, to design a computer in the days when personal computers were just in their infancy. Joe shares with us the thoughts and reasoning that went into the design of the 400 and 800, and thus brings all of us a little closer to being part of an amazing time in computer history. Please enjoy.
RK: Okay. I’m here with Joe Decuir, one of the designers of the Atari personal computer system. Joe, how are you today?
JD: I’m having a good time.
RK: Good. Can you tell us a little bit about where you currently live, and what you currently do today?
JD: I live in Issaquah, Washington. I moved from the San Francisco bay area to the Seattle area, to work for Microsoft in 1992, because the company I worked for went out of business, and I needed a job, and I dragged my wife and kids up here, and I’ve been here ever since. What I do for a living, my “day job” is, I’m what they call a standards-architect, at a company called CSR, as in Cambridge Silicon Radio. What I do for a living, is I write widely-distributed public engineering standards. I’m sure some of you’ve heard of Bluetooth, and you’ve certainly heard of USB, and I’ve worked on a whole lot of other stuff in the last 28 years. Made a difference in the world. That was fun. My main avocation is this global humanitarian technology conference that I’m going to be running in Seattle in October.
RK: What’s that about?
JD: Well, I work for… I’m a volunteer at an organization called IEEE, that used to stand for Institute of Electrical and Electronic Engineers. But it’s decided to rebrand itself as Advancing Technology for Humanity. We thought “hey there’s a lot of people out there that don’t have first world problems.” At least 4 billion people in the world live on a dollar and half a day or less. This conference is about them. It’s about rural power; it’s about rural healthcare; it’s about rural communications; it’s about disaster response, and that kind of thing. I really believe in this stuff. I had to choose between doing a good job at my day job, and doing a good job with this, I would do this. This makes more difference to more billions of people.
RK: Yeah, that definitely sounds pretty neat, I’m sure that that’s a lot of fun to work on.
JD: Yep.
RK: All right, well, I know you’re working on a book, let’s If we could, jump in to the Atari here. I know you’re working on a book, having to do with the Atari personal computer system. Could you maybe tell us a little bit about that, what that’s about, what the purpose is, that sort of thing?
JD: Yes. As a part of this I have a lot of books and other documents in my garage, and my wife would like me to get rid of them, and to do that, I first-of-all want to turn them into a book. The MIT Press, a few years back, did a terrific book on the Atari 2600, called “Racing the Beam.” It explains pretty well, how we designed it, and a lot of the cartridges that went into it. Separately a few years later, somebody wrote another great platform book from MIT Press called “The Future Was Here” about the Amiga. The book I want to write is in between. It is a sequel to “Racing the Beam,” and it is a prequel to “The Future Was Here.” The first chapter in this book will talk about what lessons we the engineers learned from the successes and failures of the 2600, and how we chose to address those in our second system. At the end of the whole process, the last chapter in the new book will be a prequel, what lessons we learned from the Atari personal computer system, and how we address these in the next system. So those starting and ending chapters will tie the three books together.
RK: So what’s your time schedule for getting that out the door, do you have one right now? Or are you just trying to work on it as you can?
JD: I have 45,000 words in the document so far. Maybe it’s 45,000 characters, whatever, it’s large, but it’s not done. There are two things I want to do, and they’re distinct. One of them is I want to find more of the original participants, and tell the story in their words, rather than just mine. The most important is a fellow named Doug Neubauer, and I really want Doug because he is still alive, unlike some of the other designers, he was the main designer of the POKEY chip, which is all the non-video I/O, and he also wrote the definitive game cartridge for this system, which was Star Raiders. If I can find him I’m great. He lives in Oregon apparently, but he hasn’t returned my phone calls. The other thing I want to do, in this book, is this machine was designed explicitly so that all kinds of people could port games and create games on it, for it. The Atari 2600 was not a development system, it was a play system. The Atari 800 system was designed so that you could write games for it, on it, and then sell them. A big chunk of the middle of this book needs to be about how you would actually take handful of popular coin-op arcade games and port them to this platform. I actually want to show the design at some high level, and some of the code. So that people can go off and write their own.
RK: Yeah that sounds very interesting, that certainly would be very different than you typically see in these types of books. So if you don’t mind, can we talk a little bit about the history, I know that you were intimately involved in the creation of the Atari personal computer line. Can we talk a little bit about what was the impetus for creating that, and why did Atari decide to go forward with creating a computer line?
JD: Well the first thing you have to understand, is we started off in the games business. We started up in the coin-op games, and then we started bringing things home. We brought Atari Pong home first in 1975. That said, we then wanted to then bring the complicated arcade games home as well, and the only question is how are we gonna do it. We designed a special and complicated system that was a line-buffer only, not a frame, and that made it achievable, financially, in 1977, when the product came out. But we knew, when we started the process, that semiconductor technology was going to catch up with us, and it was gonna become practical to design systems with frame-buffers rather than line-buffers, and a lot more horsepower, so it would be possible to bring home the things that we were pioneering in coin-op. We thought, in fact, when we started the 2600 process, we thought it would sell maybe for three years, you know like ‘79… no ’77, ’78, ’79, maybe ‘80. We knew… we thought that we had to replace the product, before our competitors did, we assumed that they would, and they certainly tried. So stepping back, we wanted to build the next great game system, and we wanted to build a computer. There was a big ongoing question about if they would be the same project, or different projects. We thought, erroneously as it turns out, that the great home computers had to play great games. The market didn’t support that. If you look at what happened in the 1980s, the Macintosh’s and the IBM PC’s and their clones, took over the personal productivity world, and the Nintendo NES’s and its successors, took over the home console business, and so they were considered separate things. In the late 1970s, we didn’t know that yet. So we tried, twice, to build a great home computer, that would also play great games. The second instance was the Amiga. Which was a monster game console, disguised as a personal computer. In retrospect I wish we had sold the Amiga as a game console, because it would have blown the doors off of the NES. The NES caught up with it with the super NES in 1992, is when they finally caught up to the Amiga and what they could do. But I’m digressing.
RK: So, yeah that’s fine. So as far as the home computer competition, though, at the time you guys decided to build a home computer, what was your biggest competition, and what did you think you could do better than the other computers that were out at the time?
JD: Those are two separate questions, and I’ll try to cover them both. First of all, I was one of many people who started working on S100 machines, in 1975 and 1976, and I was in the homebrew computer club and I hung out with those people. The first machines were the S100 machines, and then Wozniak introduced the Apple I, let’s see at April 1976, at Stanford. We thought “there are going to be personal computers, the question is what are they going to do.” The second wave of personal computers was appliances, the Apple II was the best example, the Commodore PET was less so, the TRS-80 wasn’t bad for what it did. All three were pointed at productivity applications. An old friend of mine, Bob Frankston, wrote VisiCalc for the Apple II, which was the killer application at the time. So we were thinking “we were heading in that space.” A problem that we knew were going to have, and we didn’t know the FCC was going to change the rules, we were designing for home computers… home users, not office people. So we figured that the home users weren’t going to buy separate monitors, ‘cause good color monitors didn’t exist in the late 1970s. They were precipitated to keep up with the IBM PC which came out in 1982. Right, so we’re five years before that, we don’t know this is going to happen. We designed therefore for home televisions. The 2600 was designed that way, and the 800 system was designed that way. We said, “given that we want to do personal productivity, how many characters can we put on a screen?” The physics, if you drive a home TV, you’re stuck with the upper limit of the color clock, which is 3.58 MHz. So that’s your top clock rate. So at best you can do 40 characters across, or 160 pixels across. That’s the best you can do. Does that make sense? We were driving at that. So the question is “what are we going to do that they can’t do?” Well, we’re going to be driving home TVs, which the Apple was not designed for. In fact, if you want to an Apple Store, and bought an RF adapter from a separate company, you could drive a home TV, and the FCC didn’t go after Apple, because the FCC hadn’t noticed yet. But we were Atari, and we were stuck with having to connect to TV leads, which were also indirectly connected to the antennas, that could cause interference with your neighbors, we didn’t want to do that ,the FCC would crucify us. So we designed systems that lived inside of what I would call a Faraday cage, are you familiar Faraday? Famous scientist from 200 years ago. If you basically take the circuit board, and wrap it in aluminum casting, then very little radiation gets out of that box, that’s the way their designed. Which then stuck us in a difficult problem, because we didn’t have slots, like S100 boxes or Apple II’s, that means we had to do external peripheral expansion. That means we had to make up a serial bus, the first one that I was aware of, and connect it to separate external peripherals which had their own processors in them, plus a printer, or plus I/O, or plus a disk drive, or whatever. So we were breaking ground, there. It turned out, in retrospect, the FCC didn’t drop the ball on Apple for another several years. That means that we had a price disadvantage selling as a full computer against the Apple. By the time the FCC did change the rules, we then had to face Commodore.
RK: So, can you help answer, once-and-for-all, where the names Colleen and Candy came from?
JD: Ah! They were Atari employees.
RK: They actually were? Ok.
JD: They’re now called an administrative assistant, but at that time they were secretaries, but that’s what they did.
RK: Were they picked for some particular reason or?
JD: No. Marketing told us their names, but they were employees, I’ve actually seen the badge of Colleen, she was a redhead. I think they both worked in the coin-op part of the company. They’ve, you know… women from that time tended to change their names when they got married, so it would be rather difficult to track them down.
RK: So, if you don’t mind can we talk about some the critical design decisions that were made along the way, and why those were made?
JD: That’s what I’m here for.
RK: For instance the processor. Why did you decide to go with the 6502?
JD: Ok. You probably are not a scholar of processors from that period. But until you jumped to 64-bit stuff, I think the 6502 was the best 8-bit processor at the time. It is architecturally superior to the 8080. It was architecturally superior to the 6800. I think it was still better than the Z80. It was pipelined. It was little-Endian. Which means if you’re doing… You can in one instruction… Let’s suppose you have an instruction that’s two bytes long, the first is the opcode, the second is the offset in zero-page. That instruction can point to a 16-bit number in zero-page, which is either pre-indexed by the X-register, or post indexed by the Y-register, so you can do address generation and calculation on the fly. The 8080 can’t do that, the 6800 can’t do that. Because of those architectural features, it was possible, and let me underscore, capital-P “Possible”, to build the 2600 with a line-buffer only, and not frame-buffer. Given that, we had a lot of them in-hand, they tended to run fast. We still liked them architecturally better than we liked the other processors, so we used a faster version in the computer, but we still used basically the same machines. We tried and designed a… what I would call a 6809. We did a complete design for it, we scouted it out with some of our silicon vendors, and they said “well it was extra risk, to you know, administrative risk in the design of the product, to go for a 6809 with 16-bit instructions.” So we killed that project. Later, basically the same design became what is called the 65816, which was used in the Apple IIc and IIe.
RK: Ok. The video. Now I always thought the Atari computer line was, you know, really excellent on a TV set, you could tell it was optimized for the TV.
JD: Yes it was.
RK: You talked about this a little bit already, I understand. Can you talk a little bit more about some of the design decisions that were made there?
JD: Well first of all, again, we’re split between trying to build the best game player we can build, and a home productivity thing. So as a home productivity thing, it needed have enough memory to run the code itself, and then have buffers to build the things that you were working on. For that we were basically thinking monochrome characters. So we came up with a monochrome character set that would do 40 across that would work. But having designed that, we had some other things that would be useful for other stuff. So thinking about as a game machine, as we built more and more complex VCS cartridges, we kept wishing we could do more and more complicated backgrounds. So we basically took monochrome characters 40 across, and took the same hardware, and made character… you know multicolored character maps, where there are four pixels across, and eight pixels tall, and they’re four colors. So we made up color-character mapping. Having done that, we can now build very complicated backgrounds with very small numbers of bits, because a lot of bits are being reused on the fly. Now having done that, we thought “hey”, let’s build a… we noticed that the successful VCS game designers, their processors were busy following the beam down the screen, so we decided to design a simple graphics processing unit, as part of the system, that would follow things down the screen. So instead of having a fixed-resolution bitmap, or character-map. We made it so that it changes modes, on a line or a group-basis. It could have a few lines of characters, and a few lines of bitmap, and a few more lines of characters, it could do all kinds of things, and change them on the fly. We then went further and made it possible to make them scroll horizontally and vertically. If you’ve seen the 2600 game “Adventure”, you know it moves, you know, from one frame… you go off the side, and then it jumps to the next frame. In the new machine you can smoothly scroll arbitrarily large images. It turns out, by the way, that Nintendo had to pay royalties to Atari later, ‘cause they replicated that feature in the NES.
RK: Oh really? I didn’t know that.
JD: Really, it’s not widely known; and I wouldn’t wave that in front of a Nintendo lawyer.
RK: I don’t imagine they care too much today.
JD: But it’s true. So yeah, we were trying, like I say, trying to build a machine which simultaneously can do personal productivity, and really great gameplay. There aren’t many Nintendo games that couldn’t run on a 5200, except we didn’t designed that way, and we didn’t package it that way.
RK: What about sound?
JD: Sound. Sound was a case where we didn’t do very much new design. We had two channels of sound in the 2600. We went four channels, with more frequency control, in the 800/400 system, but we basically took the old design and doubled its’ size. We didn’t put a lot of effort into sound there. We went much farther in the Amiga, but that’s a another story for another time.
RK: What about the I/O bus and the peripherals?
JD: Now that was the case where, again, we are forced by federal regulations, to minimize radio=frequency interference, so we couldn’t build… If you take an Apple II today, in its’ plastic case, it makes too much radiation; you can measure it, it’s now illegal to run an old Apple II. Just that there aren’t very many of them, so they don’t make a lot of radiation, right? So we at Atari were forced to have much lower RF output. It turns out, by-the-way, that ham radio people, love Atari 8-bit computers because they’re the quietest computer that you can buy to hook up to your six-band rig.
RK: Oh really? The 800 specifically because of the Faraday cage?
JD: Well yeah. It’s the RF-quietest computer that you can buy from that period. PCs make more radiation. Macs make more radiation. Notebooks are pretty quiet, again, but it took a long time to develop stuff like that. Anyway backing up: so we are thinking we don’t get to have slots, and that drove us crazy. What are we gonna do if we don’t have slots? Well we’re forced to maintain lower radiation to have from a low radiation serial bus to connect to peripherals. So we had to design a disk drive that had its’ own processor and it. In a peripheral adapter with serial and parallel ports with its’ own processor and it. And a printer that had its’ own processor in it. We were stuck. It worked, but it didn’t have the costs that an Apple II system where you plugged everything into the back bus. What we didn’t do is do a redesign, and incorporate all of those things internally, on an internal unexposed bus. Now in retrospect I wish we had built an 800-series machine with one built-in disk drive that was right on the bus. That would’ve had much lower costs than the version with the external disk drive. But by the time we got to those kinds of questions, I had left the company.
RK: Or if that was the intent of like, what was it called, the 1450 and the 1450 XLD?
JD: Yes. The 1450 was a great idea that was two years or three years too late. We should have brought it out in competition with the Commodore 64. In fact, if you look at the design of the Amiga 500, it’s the same design; except with a much more powerful system behind it.
RK: So what about the specialized processors that you decided to put in there?
JD: Ok. Well we had… we wanted to, like I say, support great video games, and personal productivity and adapt it to a TV screen. We split the graphics processing function into two pieces; the ANTIC chip, the name of this podcast, was the device that generated addresses, and figured out what to display, and that George’s TIA, was the device that captured the data, that was fetched by the ANTIC, and then shoved in to the TIA to display on the screen, so they were mated pair. The third part was POKEY, which was non-video I/O: that’s where we put the serial-bus I/O, that’s where we put the sound, that’s where we put keyboard, that’s where we put potentiometer-scanning. The Atari 800 has exposed, on those four ports on the front, it can handle basically two axis times four controllers. Very few people ever use that feature, but it works.
RK: Two versions of the computer. So at some point along the way, you decide you’re going to have two different computers can you talk about that for a little bit?
JD: We were carrying the schizophrenia of trying to do a productivity machine, and a video game machine, all through the process. We had designed a real computer with a full-stroke keyboard; that was the 800. Then we had designed this game console, which was going to be the 400, and then Doug Neubauer came along with the definitive video game which needed keyboard control, and that forced management to put in a membrane keyboard on the 400, so that Star Raiders was playable. Having done that, we ended up with basically a game console that can play complex stuff, and a real computer that can also play games, and internally from a software standpoint, they’re identical.
RK: As far as the design of the two different computers, I’ve heard from Curt Vendel and Marty Goldberg’s book, that the 800 is really designed like a futuristic typewriter, right, is that true?
JD: Well it looked kind of like a typewriter. Manual typewriters from the period had a keyboard similar. But so did, you know… A keyboard that’s not much different, ended up in the IBM PC. If you look at keyboard that was attached to the original Mac, it was a similar full-stroke keyboard. If you look at the... There was a company that did a computer called the “SOL” , that was an S100 box, it had a similar keyboard.
RK: I understand the 400 was actually, originally designed so that you could put it in your lap.
JD: Not really.
RK: Not really?
JD: I wouldn’t do that. Remember to play games on it, you were supposed to be using those controllers, ‘cause it used the same controllers as used on the Atari game consoles previously. In fact the same controllers work on the Atari’s; they work on all the Commodore… C64’s and related stuff; they work on the Amiga’s. Same controllers.
RK: As far as system software goes, what were your thoughts when you’re designing that?
JD: Well first of all I wasn’t one of the… I wasn’t the main designer of the system software. We hired a bunch more people to help do this. What we did, was first of all we knew that the machine was gonna ship with software, unlike the 2600 which has nothing in it. Have you ever seen the cartoon that has a Sega, and an NES and an Atari 2600 sitting there? And they’re talking about the good ol’ days, and the 2600 says “I have no memory.”
RK: I think I have seen that one, yes.
JD: It’s pretty funny, anyway it’s true. The thing had no memory. This machine had memory. The base machine had a generalized BIOS, and by BIOS do know what I mean by that? Basic Input Output System. It’s abstracted things so that you can connect the display to other things; to the disk, to the tape, to the serial ports, to the printers; it abstracted stuff, which made it easier to write…. you know, it was middleware. I’m assuming you know enough about the history of operating systems to understand what we did here?
RK: You know I do, I don’t want to assume that our listeners all do. So, yeah… if you don’t mind explaining it a little bit, that’s great.
JD: Well imagine that you build a basic input output system, which from the point of view of the application software, can write to the screen, or the display, or the keyboard, or a file on the disk, or a file on the tape, or a serial port, all in exactly the same structures. Because what operating systems do, is make it so that, from the point of view of the application software developer, you don’t need to know how the underlying stuff works. Right? Way back in the 1970s, and early 80s, things like MS-DOS and CPM were abstracting how printers work, so a word processing program could print to a lot of different things, and the end-user didn’t have to know which kind of printer it was driving. Operating systems have been abstracting everything ever since. The whole DirectX operating system extension in Windows is to make it easy to write video games, and the DirectX abstracted all the different kinds of display hardware and controls. As an aside the DirectX system was encapsulated into the first Xbox Microsoft game console. So that’s a story about abstraction; about making life easier for the applications developers, and a point I would make here is, you really want to please the end-users, that’s how you make business succeed. The applications software developers, in my opinion, are the closest proxy you get to the end-users. So if make life easier for them, you make lives easier for yourself as a hardware and system vendor. I work in the “Internet-of-Things” business today, and that’s the approach I’m still taking. Trying to harmonize the upper layers, and make it easier for the applications writers. That was probably a longer answer than you were expecting.
RK: That’s okay. So then, you outsourced the development of the Basic language right?
JD: Well, we had local people that we outsource it to, but we specified what it was, and they wrote it for us. The Basic language also had a 2K piece, which we included in the system, which was all of the floating-point and mathematics routines, so those could be used by a Basic cartridge, or the built-in Basic, or they could be used by another high level language like C, or any other language that you gonna write for this machine. Yeah, now you made a… had a comment about the right cartridge slot. We were making ROM-based systems, but we imagined that for the full computer, you might have a master cartridge of some kind, let’s suppose you have an educational cartridge, and here’s the master in the left slot, and then it might have a number of subsidiary cartridges that you would buy from the same company, and plug them into the right slot. In fact nobody ever bothered with that, that I’m aware of.
RK: There were few right cartridges ever created. I know we did that on a podcast once, and there were maybe a dozen, I mean not many.
JD: This is why... We were… everyone was trying to design to a market. They may have enough features; they may not have enough features. This was a case where we anticipated a need that didn’t show up.
RK: Sure.
JD: And you never know when that’s going to happen.
RK: That’s true. That’s true. So I guess that kind of summarizes a lot of your design decisions. How much did cost drive a lot of these decisions? Or was that… did that even factor into it?
JD: Our rule of thumb with the 2600, which we are replicating with the 800 system, was that the retail cost was probably three times the bill-of-materials. My experience since then, that’s probably still true. We were... Our goal was to build the best we could build, and then figure out how to get it to fit, rather than here’s our fixed-cost target were going design to fit into that. This is one of the reasons that the book is named “The Second System”, because in classic… If you have not yet ever read “The Mythical Man Month”, you should. There’s a chapter called “The Second System Effect.” To succeed, you have to keep focused, and ship on-time, and on-budget. Otherwise you fail, there a lot of different ways for a company to fail. But as you do all of this work, a frequent danger is that you’ll think “well, that’s a great idea but I’m going to save it for later”, “and I think I’ll save that one for later”, “and I think I’ll save that one for later”, so you get a chance to build your second system. The question is how disciplined are you at including the things that need to be included, and not including the things that didn’t. Now… If we had… If the FCC had changed the rules earlier, we would’ve built a machine, sort of like an overpowered Apple II. We would’ve built something with slots, which would’ve made all the peripherals a lot cheaper, and we would have been competing successfully with Apple, but they didn’t change the rules in time; we had to put a Faraday cage in there. Then when they did change the rules, in my opinion, Atari did not respond aggressively enough, to take advantage of the lower costs in packaging.
RK: So how much were you… I know you were involved in the development of the SIO bus. How much were you actually involved in the creation of the actual peripherals that ended up coming out for the Atari’s?
JD: Not very much, ‘cause I was up to my neck in trying to get the main chipset to work, and you know, doing all of the other software work I was doing. We did, early on, reach the conclusion that we needed a serial bus, and that I was one of the small number of people that wrote out the specifications for what would be on the serial bus. Then we incorporated some of that into the design of the POKEY and that guys like Scott Scheinman took the bus and went a lot further with it, and he was intimately involved in designing some of the peripherals. I wish.. I want to find Scott Scheinman as one of the people to include in the book.
RK: I was gonna say he would probably make a good interview, too.
JD: If you can find him. If you can find them I want to talk to him.
RK: Okay, then maybe we can help each other out with that, I don’t know.
JD: Yes, I want him and Doug Neubauer.
RK: What has been your involvement with the Atari history books that Kurt and Marty are working on?
JD: I’ve been corresponding with them for years. I have fed them a lot of information that’s been in the Atari Museum, and some of it’s been in there, ended up as photos and anecdotes in their book. I consider the Atari as you know “Atari: Business is Fun” book to be pretty much the definitive social history of working at Atari. Now they’ve been struggling to create the second book, which was “Atari Corp.: Business is War.” I didn’t have anything to do with Atari Corp.
RK: I understand you were a member of the homebrew computer club back in the day.
JD: Yes I was.
RK: What was that like?
JD: That was… oh God that was fun. Of course it was for hobbyists’, right? You know I started… I was working in medical research into late 1975, but I heard about the homebrew computer club, and I drove down a couple of times and started meeting these people. Through them I met the people who hired me at Atari, actually. Then I was still hanging around… I didn’t have a wife and kids yet… so I had discretionary time. So I was there in April 1976, when they introduced the Apple I. I think I probably stopped going there by 1977, but I did go to the West Coast Computer Faire were they introduced the Apple II.
RK: Oh that would’ve been interesting.
JD: It was, oh yeah.
RK: You still have any of materials from that show?
JD: I probably lost it all. This is some of the stuff my wife is trying to get me to get rid of from the garage.
RK: Oh really? Well if you ever... If you find any of that stuff, and you want to… One of the things were trying to do with this podcast as well, is to get things like that archived, so that they’re not lost to history.
JD: I have scanned some things. You can go on web search and find a scanned copy of my original 2600 programmer’s guide for example, you can get that. Like I used to have some of the mylar printouts that you use to make the chips, and I sent those to the History Museum, probably 10 years ago, so they are keeping track of some of that stuff. Now I have a few copies of Atari PCS and Atari VCS, and a handful of Amiga’s, because I even have children that I want to give them to. I’m in touch with the people at the… there’s… Paul Allen has a museum, where he has a lot of old computers that are still working. It’s called let’s see, what’s the name of it…
RK: The Living Computer Museum
JD: That’s it. I’ve been there. They’ve got great stuff. In fact they’ve got 2600s and 800s and 400’s and Amiga’s, still working.
RK: I mean… so you’ve met Steve Wozniak, and Steve jobs. I know that Steve Jobs had some involvement with Atari. Did you meet Steve as part of that?
JD: Yes he did. Several times. I had to turn him down several times for a job. The first was in February or 1976, which is when he snatched Rod Holt from my Atari team. Rod Holt did the design for the switching power supply in the Apple II. I turned him down again in 1977, and I turned him down again in 1979.
RK: I interviewed Al Alcorn, and he was going on-and-on about how Steve was luring people away from Atari.
JD: Right, and I stated it. I chose to work at Atari, rather than at a medical electronics company, in late 1975, because I thought that Atari would teach me a lot of stuff that I didn’t already know; whereas the electronics medical company would feast on what I already knew. I think you made the right decision. I got to design lots of software, lots of systems, lots of integrated circuits. It was a great time. I miss parts of it, but the company changed out from under me, as companies often do.
RK: How did you enjoy working at Atari, though, everybody we’ve talked to just really seem to love it?
JD: I used to go to work thinking “I can’t believe they’re paying me to do this.” I expected to come into work, and find a little, you know, coffee can with a coin slot in the top, that I should put quarters in to keep working. It was that much fun. I figured that out there in the world, there are probably hundreds of people, who would kill me to take my job, likely they didn’t know who I was.
RK: Yeah, well that’s great, that’s the kind of job you want, isn’t it? So you left you said in 1979?
JD: I left in the middle of 1979. Because some friends of mine and I had decided to start a communications-oriented company. I had been reading IEEE publications, and I had been reading about the French Mini-Tel system. You’ve probably never heard of this, but the French government was trying to reinvent telephones in the ‘70s, and they built a system with very cheap terminals, with character… characters and graphics characters, and simple keyboards, simple modems, and they built a computer system behind it. They were prototyping, the applications that made the World Wide Web explode in the 90s. So the French were almost 20 years ahead of their time. But they prototyped all this stuff, and I read all that, and I said “oh my God the future is in communications.“ So I left Atari to form a company involved in designing modems and other telephony-oriented peripherals, and ways to connect them to computers.
RK: So had the 800 and 400, they had probably just been announced?
JD: What they had been announced. I stuck around at Atari until production was working. When the production yields got high enough, is when I felt I was willing to leave. I wouldn’t leave until the production yields were good. So, in fact, as a side piece of the work I was doing in 1979, as I wrote a bunch of the system diagnostics that they would use in the factory to tell if they had a problem and where to fix it.
RK: Is that some of the information that ended up like in the field service manuals?
JD: Yes, although I don’t have the original copies of those anymore, I wish I did,
RK: Do you have any eight Atari 8-bit computers today?
JD: I have several. I have three original 800s. I have an 800 XL. I’ve got a 600 XL. I’ve got an Atari Corp. 65 XE. I have a bunch of peripherals for these things. I went out and bought, recently, a little adapter that takes SD cards, and emulates a disk drive. Because I want to use all those tools when I finish the book, because I have to prototype a bunch of software.
RK: Is it like the SIO2SD from Lotharek?
JD: Yes. Yes, that’s exactly what it is.
RK: Yeah, that’s a very popular one, I’ve got one of those as well. Those are really nice. It’s amazing how much you can improve those machines today, using something like that.
JD: If I had a magic wand, I would build… put together a little company that built a cost-reduced, and simplified, PCS system, set up to write new games on. But what’s happened is people have written emulators for PCs that will do this.
RK: Right, right.
JD: So I’m hoping that some of the audience for the book, will actually, with either an original machine or with an emulator, try creating the games that I’ll be describing; and then have fun in new ways that they wrote themselves.
RK: I guarantee you some people will do that.
JD: I don’t have a 400 thought.
RK: You don’t have a 400?
JD: Nope. I have… I wanted the real keyboards. I have been to a few classic games expos, and given some talks. I can send you some of the stuff that I’ve written. You saw the one that I gave a couple years ago in Portland.
RK: I did, yes, I watched that video.
JD: I’ve been invited to one that’s in October, in Portland this year, and I’ll probably go to be on a panel with Dave Crane, who is a great buddy.
RK: Oh excellent, that’s probably the Classic Gaming Expo.
JD: That’s the one, yeah. It’s on my calendar somewhere. It’s the week after this conference I’m running on humanitarian technology. For me that will be a vacation…
RK: So is there anything else I should have asked you Joe? Or that you’d like to say?
JD: You asked me a great bunch of questions, and I sort of don’t know what else to offer at this point. If I was retired, I would probably running some more games. Example: see, there’s a guy named Ed Friis. And he used to work at Microsoft, and he’s the original writer of the game Halo. He ported Halo back to the Atari 2600; I have a copy of that cartridge. I would like to see what happens to port Halo to the PCS. In fact, maybe I’ll ask Ed Friis permission, and maybe I’ll make that one in the book. He lives close to here; he’s a friend of mine.
RK: Excellent. Well, I didn’t really have any other questions, unless you had something else you wanted to say.
JD: Nope. If there’s nothing else I can do, you can correspond with me later. Thank you very much Randy.
RK: Well thank you very much for your time I really, really appreciate it. I think this is going to be very interesting for everybody, and you have been excellent.
JD: If you get more questions, you can forward them to me, and I’ll answer them as best I can. It’s a promise
RK: Ok. Very good.
JD: Cheers dude.
RK: Thanks again. Bye.