Posted on June 16, 2017
Chris Crawford’s Famous Dragon Speech and Interview From 2008
Chris Crawford recently highlighted a “cleaned-up” version of his famous, game industry-defining “Dragon Speech” from GDC 1992. In this speech, Chris Crawford explained his dream of of true “interactivity” and how the game industry and he had parted-ways. While he was almost universally rejected at the time, in an era of machine learning, A.I. chatbots, art-games, indie-games, and serious games, Crawford’s ideas about the true purpose and goal of computer could not have been more prophetic. Crawford is my true hero, and I’m so happy to get to see this speech for the first time.
Also, below is an interview I conducted with Crawford about his time with Atari from 2007, published in 2008.
-Steve Fulton (Fultonbot)
(The following interview was published Dec. 22, 2008)
Chris Crawford was hired by Atari in 1979 as a VCS programmer. He soon moved to the 8-bit computer line where he programmed one of the influential games of the 8-bit computer era, Eastern Front. For the past 25 years he has worked as a game designer, software evangelist, and has been a pioneer in the area of interactive fiction.
Steve: Can You Describe VCS Development in 1979?
Chris Crawford: It was a very difficult machine to program. You had 128 bytes of RAM, and 2K of ROM space and the video display was driven by the CPU, the 6502 and so basically, most of your code consisted of the drawing code. You drew it one scan-line at a time. Basically you’d frantically load-up the display registers with the display data for one scan line, and then you had to load up the registers with the display data for the next scan-line and you had exactly 76 machine cycles in which to do this..or on average about 35 assembly language commands. That’s pretty tight restrictions.
Steve: You did Wizard right, and that never came out?
Steve: How did you get into the 400 and 800?
Chris Crawford: It’s funny, at that time everybody wanted to work on the 400 and 800 because it was so much sexier and powerful.
Steve: It was the 6502, but was the VCS some lesser version of the 6502?
Chris Crawford: No, exact same processor.
Steve: So just less support, memory, chips, etc?
Chris Crawford: Yeah. The 6502 in the 800 was faster. They clocked it at 1.8 MHz, whereas Stella’s was 1 MHz. But it had much better video. there was a graphics processor named Antic and Antic handled all of the graphics work ,whereas with Stella the CPU spent most of its time drawing the screen. with Colleen you simply set-up a page display and let that run. There was another processor called Antic and Antic did all the work that the 6502 did in Stella.
Steve: So with Stella you have 1/2 the Mhz and no co-processor?
Chris Crawford: Right, you did not have anywhere near as many CPU cycles to play with. The other thing of course was that Colleen had a lot of memory. the smallest was 8K (as opposed to 128 bytes) and it had a big ROM with all sorts of operating system stuff in it, interfaces for nonvolatile memory and so-forth.
Steve: Did you ever talk to Nolan Bushnell while you were there?
Chris Crawford: No, I never did. The first time I met Nolan Bushnell was, God, years later, I ran into him at a little conference of techies…I forget which one it was.
Steve: When I talked to him he was very much of the opinion that the Atari that he had started was very much based on game design and making games and dealing in that realm and when the Warner guys took over they really had no idea about that business and ran it into the ground. Did you see anything like that when you were there?
Chris Crawford: I think that is partly true. Now, I’m only replaying the scuttlebutt that I heard while I was at Atari, but the story that ran around the programmers (who were fairly disinterested observers I think) was that the VCS initially did very badly and after a time Nolan felt it was time to give-up on the VCS and build something new. He was especially enamored of the home computer. It was such better technology and so his attitude was ‘dump this VCS loser and let’s put all of our money on the home computer’ and the Warner people disagreed. It was Manny Gerard at Warner, the main guy, said ‘no, we just need to develop the market some more, we need more games, we need to build-up a bigger software library, we need to give this product time’ and so the Warner people refused to abandon the VCS. That was, according to the scuttlebutt, the reason for Nolan to leave.
Steve:From your opinion, from being there, what do you think? The VCS was successful for a couple years, but then its limitations were really what made it die. Do you think Nolan was right, or the guys are Warner were right?
Chris Crawford: The guys at Warner were proved to right because the VCS did not peak until 1982, and Nolan left in ’79, so the growth curve continued up steeply in ’79. ’80, ’81 and in fact, what brought Atari down was the E.T. cartridge in Christmas of ’82. so even ’82 was a magnificent year for Atari and most of Atari’s profits came from the VCS, not the home computers and not the coin-op machines.
Steve: When you worked in the Home Computer Division, do you remember a time that it was ever profitable?
Chris Crawford: I wouldn’t know the answer but my impression was that they were always spending more money than they were taking in. The home computer did grow and it did enjoy good sales, it was doing well, but they kept adding to the home computer division, investing in it the same way they had done with the VCS, but then the Commodore 64 pulled the rug out from underneath the home computer.
Steve: The pricing rug? They pretty much cut the price in half to being with.
Chris Crawford: Yeah, there was a price-war. At that time when the the Commodore 64 came out there were a number of color computers. There was the Apple, the Atari, Texas Instruments had a machine, Radio Shack had a machine, and there were a couple of other real minor ones. The Commodore 64 came out and it was priced below everybody else, and that forced Atari to drop its price. Basically, Jack Tramiel was moving all of his production overseas, and he was able to lower his prices. There was a steady price-war, and over the period a of months the prices kept going down and down and down. What really killed Atari was they decided to move all the production to Hong Kong. The christmas production was supposed to (christmas was big selling time for these machines),the Hong Kong unit was supposed to come up in August ’83.F
Steve: For the XL line?
Chris Crawford: Yeah. They shut-down production in the States expecting the Hong Kong production come on-stream, and the Hong Kong line had problems and didn’t come online until November. When Christmas came there were no Atari computers on the shelves.
Steve: I can attest to that. Christmas ’83 I asked my dad to but me one and I ended up with an Atari 800 instead, which I loved because it was superior. I certainly remember that Christmas.
Chris Crawford: Yeah, it was an absolute disaster, a catastrophe for Atari and that is what sealed atari’s fate. Now there a bunch of other things that greatly contributed to it, but I feel that was the knock-out blow.
Steve: When you were there you designed some of the early games. Didn’t you say you worked in a research lab?
Chris Crawford: Yeah. I think had 4 different jobs at Atari. My first job was as a Stella programmer and that lasted 3 months and I wrote one program for Stella named Wizard. After that I was transferred to the Home computer Applications group where I was programming the home computer, and that lasted about 10 months.
Steve: Is that where you made Energy Czar?
Chris Crawford: Energy Czar and SCRAM were made during that time. Then I was promoted to supervisor of the software support group. Our job was to provide technical support to outside programmers. We had a whole package of goodies we provided for free. By the way, the main thing we did was this tour where I would travel around to cities all over the country. We would rent a hotel meeting room, and people could come-in to these seminars where we taught them all about how to program the Atari and I did almost all the work here. I had a real barnstorming style. My job was to wean people away from the Apple to the Atari. I was pushing that line really hard. Somebody in one of the magazines that had come to it said ‘Crawford does a show like an old-time evangelist. You half-way expect him to start quoting the bible’. and that is where the term ‘software evangelist” arose.
Chris Crawford: Yeah, I was the first.
Steve: So you moved on an did Eastern Front which was a huge success?
Chris Crawford: Yes, yes. Although, an interesting point I’ll make, Eastern Front was the classic example of technological opportunism. The way many games are designed nowadays…and I’m very critical of technological opportunism. I did it back then. The Atari 800 had this wonderful scrolling capability. I developed a little scrolling map thing just to show off to people this wonderful capability. I remember telling Joel Billings at SSI and number of other people ‘boy think of the war game you could make with this’….and they said ‘yeah, yeah, yeah’
Steve: You had a history of building war games, right?
Chris Crawford: Yeah, but I hadn’t done any on the Atari.
Steve: You did some for Avalon Hill for the Apple?
Chris Crawford: Actually, I did them on the Atari and people ported them over. I did Eastern Front on the 800 initially as a demo, and when I tried to interest war game people into using it, they kind of blew me off. So on the side, on my own, nights and weekends I said ‘well, let’s see if we can add some units here and move them around. Basically one thing led to another and I ended-up with a war game. I will tell you an important point to make for the readers is that game was TOTAL CRAP before it shipped. The game play was atrocious. It was really dull and boring and I had the good sense to realize ‘ship it when it is ready’ so I just went back to the drawing board and reconsidered how it was working and made some major changes in its operation. It worked!
Steve: They allowed you to do that at the time?
Chris Crawford: I was doing it on my own, that was the whole deal. If I had been doing this as an official project it probably would have shipped before it was ready.
Steve: Eastern Front went out via APX instead of through the Atari software channel?
Chris Crawford: I showed it to the Atari people, the marketing people, and they said ‘oh geez, this will never sell. it’s a war game’ they said ‘you can put it in the Atari Program Exchange”. I put it there and it was a huge hit. The next year they came and said ‘why don’t you do a new version for us that we will release as an official Atari product?” So you know, they were just completely wrong.
Steve: How successful of a product was it for APX?
Chris Crawford: Yeah, it was really the product that made APX. That along with Caverns Of Mars. Those two products together made APX a huge success. By the way, there is a side-story on APX. The guy who cooked up the idea, Dale Yaokum, was trying to explain to the management that there are a lot people out there that like to write programs and if we can publish these programs for them, it’s a win-win. The management was not very interested in it. He put together a business plan for it and said ‘look, we only need a little bit of money and this thing can be self sufficient and it might make some money.’ They very grudgingly agreed to let him do it. And so he did it and very quickly made it into a monster success. It was a major profit center for Atari. They rewarded dale for his initiative by bringing in another guy to be Dale’s boss and the other guy didn’t know anything about software! The other guy was really hard to work for, so Dale in disgust, quit about a year later. Classic story of executive blunders.
Steve: What did Dale go on to do?
Chris Crawford: He quit APEX and went over to corporate research. He ran a unit that was going to design a “shoot for the moon” new computer. The worked on something with a 286, the absolute newest processor coming out and they were getting pre-samples from Intel. They were were designing that when Atari collapsed. He then got a job at Xerox PARC, then founded his own company. About 10 years ago he sold his company for millions of $$, retired, bought himself an airplane, learned how to fly…he always wanted to be a pilot, and he’s now very happily retired.
Steve: In the beginning, did Atari management want to own all the software for the computers?
Chris Crawford: Yes. The attitude of the executives was ‘we want to make all the money on the software. We don’t want any competitors. They were having competitors with the VCS and the programmers were trying to explain that ‘no that’s not how it works, you need a big library of software, you need to encourage them’ and I was one of the people doing that. Initially they had never quite defined what it was that had to be kept secret. I was the programmer at Atari who had come-in from the outside world and had more contacts with outsiders. I’d be working on Atari software and the phone would ring and it was somebody in Indiana saying ‘can I get any of the technical documents?’ and I would go over to the main area and get a few of the technical documents, photo-copy them and mail them off. I was sending out…there were enough loopholes that I was able to send out some documents and not get fired.
Steve: But for the most-part, they wanted to keep a lid on all that documentation?
Chris Crawford: Yeah, they wanted it all kept secret. I was sending out some minor stuff and then one day it was sort of like ‘the dam broke’ and they had an officially policy, 180 degree reversal ‘we want to tell everybody about this’. I immediately got on the phone and started calling a bunch of my contacts saying ‘hey would you like complete technical documentation on the Atari?’ and we shipped a lot of those.
Steve: What did you think of M.U.L.E.?
Chris Crawford: Yeah, and I think this opinion is shared by most of the designers who were active at that time. My belief is that M.U.L.E. was the finest computer game design ever done in terms of the going “with the grain’ of the machine. Using the machine to fullest, really understanding what the machine could do. It was just a beautiful design because it was so perfect for the Atari.
Steve: The use of joystick ports, the sound?
Chris Crawford: Not just technical brilliance, design brilliance too. He didn’t use a lot of techie tricks, he actually used them in a very creative way and himself made a game that was brilliant.
Steve: It’s been 15 years since you’ve released a game and you are now finishing up Storytron interactive fiction engine which was created out of the relationships in Excalibur in 1983, What is the breakthrough that got Storytron ready to go in 2007?
Chris Crawford: There were many breakthroughs of major developments that I’ve had to make, and that is one reason why it has taken 15 years. If it was just one technology I had to build then it would have been done years ago. moreover, there is a strong synergistic relationship between these components and so I could not see them all at once. I started work on the basic engine and it in itself was a breakthrough in how it handles drama and so-forth, and it was only one and by itself it was insufficient. I didn’t realize that it’s biggest problem was that it was very difficult to program then engine, to give the engine the data it needed to tell good stories. That was the engine I did between ’91 and about ’94 and ’95. The next big breakthrough was building the editor that allowed a user to program the engine…to develop the data set required for the engine. It’s a very complex data set and it took me a year to that because I had to build a scripting language and the fundamental requirement was…if I was just writing a scripting language for programmers it would have been trivial, I could have knocked it off in a few months, because everybody has done that, but my requirement was that this had to be something accessible to non-technical people.
Steve: So someone like novelist could sit down and use this?
Chris Crawford: Right, but it was still programming and so it’s still la huge restriction on the novelist, we are still demanding an enormous amount from him. We wanted to eliminate all unnecessary techiness and that was a huge task. I did a first round on it in period of ’94-’98 and it was no where near good enough. It was functional, you could do things with it, but it was still very difficult for people to do things with it. this was the “erazmatron” period. For Storytron I tore it apart and completely re-built the entire thing from scratch and came-up with an even better scripting system. It’s still difficult to use. This is like Macromedia Flash. This is non-trivial., but it is also a hell of a lot easier to use than a real programming language.
The 3rd killer problem was the user interface. In erazamatron the user interface, basic internal structure, the basic atom was a ‘sentence’ and you interacted with people one sentence at a time. That was pretty limiting and it really did constrain the designer rather badly. I’d say the biggest of the breakthroughs was the linguistic system I have in Storytron. Basically it allows the user to speak to the computer in this toy language. It’s a very powerful language system. The Story Builder creates words and defines them, and that is process of creating a work in Storytron. You define all the words…of course, defining them is a big job. That’s the concept.
Steve: I’ve read that your plan with Storytron, and this might have been someone’s idea of what your plan was, was to create a kind of a myspace like web site where people could come and create their interactive stories and share them with other people. Was that your idea, or someone else’s?
Chris Crawford: That’s not quite what we are doing. what we have is web site where, when it is fully operational (we are planning for January 1 (2008) and we are on-track) basically anybody who wants to can download the authoring tool and use the authoring tool to create a story world. When they have the story world the way they like it, then they can upload it to us to put into our library. Then any consumer can come along and play story worlds in the library. Our revenue model is that we make money from the consumers playing the story worlds. Revenue is hared with the authors, and wearer aiming for a 50/50 cut.
Steve: Do people subscribe or do they pay per story world?
Chris Crawford: Initially it looks as if we will do pay to play, or fixed fee where you get to play one story world per month or two months or something. Once we have big library we will open it up on subscription basis, however I will say we have not ruled out an ad-based revenue model. We simply had to choose one or the other for our business plan and we felt the subscription based had some advantages and we went with that, but we may end-up ad based. We don’t know.
Steve: What’s encouraging is that you are actually ready to launch this.
Chris Crawford: Yeah, this thing is definitely coming together.
Steve: So, have you heard in the past anyone describe you as a Don Quixote like character?
Chris Crawford: Oh yeah, I have!
Steve: What do you have to say about that now, now that you are almost finished?
Chris Crawford: You know, they’ll see. In fact, I’d like to address the criticism you mentioned ‘Crawford hadn’t designed a game in 15 years’. The fact is, I have not made any effort what-so-ever to talk to people about games in 15 years. Every single public press…i mean the books i’ve published, I’ve published 3 books, one was on Interactivity, one was on Interactive Storytelling, and one was about games, and the publisher approached me, and said ‘geez, we’d really like you to write this book’.
Steve: And that book ‘On Game Design’ pretty much covers all the games you wrote. Period.
Chris Crawford: Yeah, I make no attempt to teach about the current generation of games. All of my public presentations have been at the request of the host. I’ve never gone out and looked for it, they just call me up and they want me to talk. In a couple cases I’ve told them, ‘I haven’t done a game in 15 years’, an they say ‘well,we still want you’
Steve:You are going to be done soon, do you have anything else to say before we sign off?
Chris Crawford: Well, the Storyton stuff is definitely going to change things. The games industry has gotten stuck in a rut doing the same things over and over again, making fundamental errors a long time ago that are now holding them back. some critical mistakes. I think the first mistake was around 1990 with Wing Commander. Wing commander was very bad for the industry because they bought market-share. They threw an awful lot of money at that game and produced a game that was very expensive. The game never made money, the add-on packets brought them into the black. They did the same thing with each of the subsequent Wing Commander games. The basic game itself ended-up losing money because they spent so much money on the graphics as a result the games industry is now very capital intensive. They send millions of dollars making a game and there is no easy way to build a good game that can get a fair shot in the marketplace. That means they have cut-out one of their best sources of creative input which is all the crazy people out there. the model I like to use for this is, Hollywood has it nailed down and the games industry really should learn from Hollywood here, although it night be too late. Basically there are 8 million people (surveys show) 8 million people in this country will tell pollsters ‘yes, I have an idea for a novel I want to write”. Out of those 8 million a few hundred thousand apparently, each year actually write something. Out of those few hundred thousand, I think it is something like 10,000 actually produce a manuscript that they ship to a publisher. Out of those 10,000 only a few hundred are published. Out of those few hundred, only a handful actually hit the big time. Maybe a dozen make a goodly amount of money. Out of those dozen, 1 or 2 will be cherry picked to make a movie. Think of it as a pyramid creative base is 8,000,000 ideas and at each level there is a selection that takes place that knocks out 98% of everything. It’s a sorting system that takes the very best for the full treatment.
Steve: And you are saying the games industry really does not have that because no small person cam sit-down and write a full-fledged game without $10,000,000 or more?
Chris Crawford: Right now anyone in this country with a word processor can sit-down and write novel and it might be a huge hit.
Steve: Do you think stuff like Storytron (and other technologies) are sort of changing that for small part of games industry?
Chris Crawford: Well ,that is certainly what we are doing with Storytron. We are using the Hollywood model, not the games industry model. My point here I suppose is that this is one of the greatest weaknesses of the games industry. they just can’t tap-in to this huge base of creativity. Yes, there are lots of tools that allow you to build interactive fiction or platform games or first person shooters or so-forth, but the problem is, they are starting off with the assumption that you are doing a genre.
Steve: You can’t create an interesting game, because the game has already been created for you. you are just editing the game.
Chris Crawford: Yeah. Where did J.K. Rowling come from? She just came out of the woodwork. Same thing with Tom clancy. Classic example of an absolute nobody who had the right combination of talents, he slapped together books and kaboom! The system really worked for him. It picked this guy out of obscurity and generated millions and millions of dollars of wealth. the games industry can’t do that.
Steve: So that is really the story of Storytron then? A model that is separate from the games industry to open something up?
Chris Crawford. Yes we have no desire to compete with the games industry on anything. We are a completely separate market.
Steve: Do you think the games industry has a narrow view of what can be called a ‘game’?
Chris Crawford: Yeah, in the 80’s when I was working in it my feeling was ‘let’s expand this definition to include all sorts of crazy things, let’s make this a medium of expression’ and they narrowed it down to ‘a hobby’ which is not a medium of expression it is a hobby greatly by a small collection of people.
Steve: Do you follow the Nintendo Wii and what’s been happening with that?
Chris Crawford: I’m aware of the Wii and the fact it has generated so much excitement and the fact that it really is a radical new concept for the games industry.
Steve: Well it is a radical new concept and I love it myself for those exact reasons, but what I find most interesting about it is the reaction from the “hobbyists” that you described. Almost a fearful reaction that somehow their hobby will be destroyed by something new, some new expanded market, and it might be exactly what you are describing.
Chris Crawford: I was unaware that the old timers were uncomfortable with the Wii.
Steve: Yes, the hardcore gamers are uncomfortable that the Wii is taking market share away from PS3, the Xbox360 and the PC, and the games are more oriented toward multi-player and family oriented stuff. their worry is that it is going to destroy their hobby, which I find quite funny myself.
Chris Crawford: Well there is actually in one sense I suspect the Wii is a continuation of an old evolutionary process that has been going on in all forms of software. The basic sequence here is that you get a piece of software, or a game or whatever. It’s successful and there are a bunch of people who really love it, so you come out with version 2. The thing is, that sells primarily to the aficionados who loved version 1. the want something better than version 1, meaning something more complicated than version 1. Version 2 always had more features than version 1. This process continues with version 3,4,5,6 etc. The problem is, by the time you get to version 5, you have built something so hairy that the average beginner can’t use it. At that point, somebody else comes in with the new easy to use version, a a clean simple one, and that attracts a new generation of people who are intimidated by the monster version the aficionados like, and this process just keeps going on over and over. The Wii in a sense is something like this. My impression from the software available for it is that it tends to be beginner-level software.
Steve: The games on other systems look so complicated or you need to play online you need to be yelling at people through your microphone. A lot of people don’t want to do that.
Chris Crawford: Yeah, it’s a regular cycle that software goes through and I fear…I don’t know how to stop the cycle. Our work is going to get more and more complicated and at some point will we ebnd up being replaced by someone who is cleaner and easier? It looks like…if you are the guy who owns version 5, you are not going to throw it away, you are stuck with it. It’s a tough problem.
Steve: Do you think the web is your next direction for software delivery?
Chris Crawford: There is no question that the web is the future. That’s sort of obvious I feel. The web is steadily taking over everything. I’ll mention one advantage of the web. Our software is…we don’t have to worry about piracy at all. The basic engine that runs everything never runs our web site.
Steve:Right, it can’t because it’s tied to the back-end
Chris Crawford: Yep.
Steve: You give the Story building engine away for free. Piracy is nothing to you.
Chris Crawford: Yeah, it’s a really nice model. I just can’t understand why there are people who haven’t just abandoned everything else.
Note: This interview with Chris Crawford was conducted in September on 2007 as source for my Atari History articles I was (still am?) writting for Gamasutra.com. Since this interview was conducted, Chris Crawford’s Storytron site has launched, and both the development tools and the player are now available for download.