Interview with David Joiner (Talin)

It's a "handle" I started using at science fiction conventions, and I just got used to people calling me that. I liked it better than my birth name.

I actually learned programming in the U.S. Air Force training academy. They taught assembler, COBOL and Fortran. After that, I spent 4 years (1976-1980) doing COBOL programming at the Strategic Air Command headquarters in Omaha, Nebraska. The programming I did there was logistics database support - for example, failure analysis reports on B-52 bomber components.

During my first few years the programming that I did for the Air Force was on punch cards, but eventually we got CRT terminals that hooked up to the Honeywell mainframe that was used by our department. The terminals came with a BASIC interpreter, and so I secretly worked a primitive "space war" game whenever I got bored with my real work. Although I could have gotten in very serious trouble if anyone found out about this (recall that this was in a high-security area in the middle of SAC headquarters!), I eventually showed my work to a co-worker - and the next week I discovered to my surprise and horror that several high-level officers were actually *playing* my space war game! Fortunately, there was no way to trace it back to me, so I never did get in trouble.

During that period I met several friends in my barracks (these "barracks" were much like a college dorm) who were computer enthusiasts and had bought personal computers such as the Commodore PET and the Apple II, and I started learning to program them. I remember a two-player, turn based tank game that I wrote in PET Basic which was quite popular in my circle of friends. Eventually, I bought my own first computer, which was an Exidy Sorcerer - a decision which turned out to be a mistake, as it never got popular, and had a number of serious design flaws.

After I got out of the Air Force, I wrote several games on the Radio Shack Color Computer in assembly language - the first one being Guardian, which was a Defender-style arcade game that was sold by Quasar Animations. That was also my first (but not last) experience with unethical business partners / employers - I think I made about a total of $300 from it. Then in 1984 I got hired at DataSoft, and started working next to experienced game programmers, and as a result of this I picked up a lot of knowledge about the techniques of game programming.

Pretty much all my programming knowledge is self-taught - I've never taken any kind of college computer science course. I learned the C language by typing in the source code to a C compiler from the pages of Dr. Dobb's.

Actually, at the time I couldn't afford to buy one, as I had just been laid off from DataSoft when they went bankrupt and were acquired by IntelliCreations, which shortly thereafter went bankrupt and was acquired by Mindscape. At the time I was sleeping on various people's couches, and doing odd bits of contract work to pay for things like food.

I did have access to a pre-release Amiga 1000 system owned by my friend Ken Jordan that he'd gotten by signing up as an Amiga developer. I started exploring the machine and what it could do.

In fact, I first heard of the Amiga from Ken while we were both sitting in the lobby of DataSoft waiting for our job interviews to be hired there. I was telling him that I had heard about this new Atari ST computer (the "Jackintosh") that was coming out, and he set me straight and told me that there was a much better computer coming out called the Amiga. We quickly became friends, and had lots of conversations speculating about what the Amiga would be like, and the kinds of things we wanted to create for it.

It wasn't too bad actually. We had early versions of the Amiga ROM Kernel reference manual, and it wasn't too hard to puzzle out how to get things to work. In many ways, it was simpler than programming on the Apple II or the Exidy.

So MicroIllusions started out as a spin-off of KJ Computers, which was a computer store in the San Fernando valley. I used to go there and hang out a lot, even though I couldn't actually afford to buy anything. One day the store's owner, Jim Steinert, asked me if I could make the machine talk - supposedly the Amiga had speech synthesis and he had never heard it. I asked if I could borrow a machine and the manuals for a few hours, and I was able to quickly hack together a program in AmigaBASIC that used the built-in speech synthesis to speak any sentence that you typed in. He was so impressed by this, he asked me to polish it up so that he could start selling it - "I'll give you half the money". The program was called "Talk to Me" and he started selling it in ziploc bags in his store for 30 buck a pop.

At that point, he realized that he could make a lot of money going into the software development business, so he started a new company, which initially was run out of the back of the store, but soon moved to an office building a few blocks down the street in Granada Hills. There was also a house next door that was owned by the same company that owned the office building, so Jim sub-rented the house to myself and my three friends - Allison Hershey, and Greg & Bonnie Long. All four of us were devoted science fiction fans, and all of us worked for MicroIllusions in some capacity or other. The garage of the house was taken over for the equipment needed to duplicate the disks (That was before we started using commercial duplication companies).

So basically even though I was "working from home", the office was right next door. It really was an exciting time for me, probably the best time of my life. We really felt like we were building the future.

That being said, Jim never wanted to pay social security taxes or anything like that, so we were always contractors and not employees whenever he felt he could get away with it. I was not paid a salary, I was paid an advance and royalty payments, and I licensed the copyrights of my work to MI. The contracts of course were horribly written since neither Jim nor I had any experience in this area, although I did get some legal advice which removed some of the worst parts of the contract.

After the speech program, Jim came to me and said he wanted me to make a program "Just like 'Cave of the Word Wizard'" which was on the C-64. I didn't want to do an exact clone, so I changed it so that instead of a little boy lost in a cave, it was a little boy lost in a crashed spaceship. I also didn't want it to be just about spelling, so I made the game extensible to other kinds of questions and puzzles.

Doing an "Edu-tainment" game was a smart move for Jim, since these products tend to be "evergreen" - that is, they don't have a big sales spike at the beginning like a regular game, but they can stay on the shelves for years - Discovery was still making decent royalties for me even 5 years later.

Discovery took 4 months to develop. I did all of the artwork in Deluxe Paint. (I've always been something of an artist, with sketchbooks full of designs for spaceships, aliens, costumes, and so on.) For the music, I created a very simple editor program that allowed me to sketch out waveforms and place notes on a grid, which then was fed into the Amiga's audio libraries.

After Discovery was finished, I immediately set to work on Faery Tale Adventure, which was a game I had been wanting to do for several years - Jim was now confident enough in my skills that he was willing to front me an advance for the work.

I think that they were spelling, math, and flags of the world. My friend Allison did most of the work on the flags one, painting them in Deluxe Paint.

There were 8 questions total, 3 of which were chosen randomly each time. Some of them were:

Make haste, but take heed.
Scorn murderous deed.
Hold fast to your creed.
Wing forth in flight.

...I don't remember the others. I do apologize for the awkward copy protection, but at the time it was pretty much a requirement. At least it was better than "go to manual page 52, line 5, word 11" which was the norm for a lot of games at that time.

Well, I would probably be about a year after the Amiga was released - Discovery took 4 months, and FTA 7 months. So it couldn't have been one of the first - I remember playing Archon before Discovery was even started.

Unfortunately, I don't remember well enough to do so.

So the word "Fairy" has a bunch of different archaic spellings: Fairy, Fairie, Faerie, Faery, and so on. I picked one of the more obscure ones, pretty much for the reasons you suggest.

At the time it was fairly common for a single author to do an entire game, but I had an advantage in being somewhat of a "renaissance man". It's ironic, because when I was growing I was never able to focus on one single creative outlet and exclude the others - and this was considered a disadvantage. People would tell me I had to learn to focus on one thing, otherwise I would never be successful, just be a dilettante. I struggled to find a profession which would use all of my skills, not just some of them - at first I thought that perhaps I would be a filmmaker, but my short stint in Hollywood (1983-1984) quickly cured me of that ambition.

It wasn't so much a matter of keeping creative control, it was more that I was developing the game on my own and I didn't know anyone else who could do the animation and music that I needed.

What's fascinating to me is that we've now come full circle - with the advent of mobile gaming, it is once again the case that a single person can create a successful game.

It took 7 months to develop the Amiga version and I was the only programmer. However, other programmers worked on the Sega and DOS versions of the game, but those were only started after I finished the Amiga version.

It was written in Manx C.

It was an improved version of the music editor I used for Discovery.

Well, one thing I realized is that people would get bored if I always played the same piece of music over and over, so it seemed fairly natural to change the music based on events within the game. At the time I didn't think I was doing anything special or innovative.

There were two editors I created: "anima" was the animation editor, and allowed me to paint characters and flip between frames using the cursor keys. The other editor was "tiled", which allowed me to edit the tiles and maps. I think that over the years, I have written 6 or 7 tile editors, including the one I am writing right now which uses OpenGL ES for rendering and runs on Android.

From the very first, I had two key goals in FTA:

-- to have a game which was based on fairy tales rather than the swords & sorcery / high fantasy genres which were extremely popular then. (I've always liked busting cliches).

-- to have a very large and seamless world, instead of being divided into "zones" where the game would pause when loading a new region.

BTW, one thing that Ultima VII did better than me was sound - such as having the character's footsteps sound different based on the terrain type he was walking on. As Chris Clairemont used to say "Sound is your secret weapon". I have never forgotten this lesson.

Lots, too many to list. For example, in order to load the map data in the background, I had to go direct to the disk driver and load raw sectors off of the floppy, instead of using files. That's why the FTA floppy disk has such a weird sector layout - part of the disk is a regular AmigaDOS file system (to hold the game program and some data files), and the rest is raw data written to the disk.

Yes - and I think I still may have that sketchbook page somewhere in my files.

I think I mostly made it up as I went along. In this, I think I was inspired by Jon Van Caneghem's approach to making Might and Magic, which was to start with a basic engine and then add detail like crazy.

It's interesting too, because many years later in working on Sim City 4 and Sims 2 at Maxis, I ran into the same principle, which is this: There's really no way to measure how much fun a game will be until you've built it, or at least built enough of it that you can start playing it. There's no way to predict the level of fun if all you have is an abstract design - partly because it's hard to separate the fun and excitement of designing the game from the fun of playing it. So the best you can do, other that copying an old design, is to build the game and then iterate on the design until it achieves the fun level you want.

Well, it kind of grew out of the tools I had developed for Discovery and FTA. Also, I had noticed that the Atari ST seemed to be taking the lead in music software, and I felt that the Amiga, which was a clearly superior platform, should have a better suite of tools for musicians.

I had also been interested in Moog synthesizers and electronic music since high school, and I had built many of the electronic music kits that were sold by Paia Electronics (which still exists, amazingly). So I understood the basic principles of additive and subtractive synthesis, oscillators and filters, and so on. I wasn't a great musician, but I had taken organ lessons as a child and so I had enough musical theory not to embarrass myself.

Another source of inspiration came from, of all things, DynamicCAD, which as you remember was another MicroIllusions project which failed badly. During this period Jim often relied on my advice in deciding what projects MI should work on, and while I didn't actually write any code for DynamicCAD, I had done a lot of the due diligence that was involved in starting the project. You see, just about everything creative I've ever done has been the result of what I call "a collision between skill sets". In this case, I took some of the ideas from the CAD world, like grid snap, direct manipulation on a grid, and so on, and applied them to musical notes.

You know how you tend to clearly remember events which are associated with strong emotions? Well, I clearly remember the total panic I was in - you see, we had forgotten to bring the power cord for the Roland D-50 keyboard and I had no way to do my demo if my MIDI instrument wasn't working. What I ended up doing was unscrewing the power socket on the back of the keyboard and using a pair of alligator clips to splice the power connections to another power cable that they had lying around the studio.

Other than that I can't recall much about it, except that it was my standard convention-floor demo.

Yeah. At the time I was trying to disassociate myself from MicroIllusions, and so I started publishing under that name. However, it was just me. I think I did get an actual business license, but I never did anything with it. I also liked the cognitive dissonance of combining the words "sylvan" and "technical". Also, I was going through an "elf phase".

My financial relationship with MI was long and complicated. Jim wasn't a very good businessman - but that was not unusual for the software industry at the time, there was so much wide-open opportunity that any half-competent person could start a software business and be moderately successful.

Jim and I also differed in our approach to business ethics. He imagined himself to be a sharp dealer, and once boasted to me how he "saved money" in dealing with the disk duplication companies. You see, at the time there were companies which would do all the duplication work - that is, make copies of the floppy disks, print the packaging and assemble the boxes. Any many of these companies offered 90-day terms - that is, you didn't have to pay for 90 days, so you could use the money you made selling the product to pay back the duplicators. This made it possible to be an entrepreneur with very little startup capital, other than the sweat equity of writing the software.

Well, Jim's idea was that when the 90-days came up, you simply refuse to pay - and then eight months later when the duplicators eventually get around to suing you, you settle out of court for like 1/3 of the money. This same kind of playing fast and loose with the rules is what caused him to lose the Hasbro contract, which up to that point had been an incredibly valuable asset to the company.

Many years later, I went over all of the royalty statements that I had gotten from MI, and discovered that there were lots of basic arithmetic errors in them - and not always in Jim's favor.

In any case, I was emotionally torn - I wanted to be loyal to Jim because of the many opportunities he had given me, yet I was becoming increasingly uncomfortable working with him. So I had pretty much separated myself from MI before the bankruptcy. At the same time, however, part of the reason for that bankruptcy was that Jim had alienated a lot of his best talent.

It's the other way around - the Installer was actually written specifically for Commodore - my partner Joe Pearce and myself did a number of small contracts for Commodore, including the installer, the CD-TV welcome disc, and the Commodore MIDI library. We used it on Music-X 1.1 later.

At the time, Hollyware was in arrears for Music-X royalty payments to me. I told them I would forgive the missed payments and give him a 2.0 version of Music-X (which I had already written, with Joe's help) in exchange for the rights to FTA. (Although they were supposed to pay me royalties going forward, which never happened. They claim it's because they never sold any copies.)

I have never seen this game, sorry.

Absolutely. Robert McNally approached me about the idea of starting our own studio - both of us had suffered in the past in our deals with development studios, and the time seemed right to start our own company.

From the very start, we set out to create an organization where even the lowest employee had some influence in running the company. And the result was that we created a company of people who were amazingly dedicated to the organization, even though we couldn't afford to pay people very well.

Our worst mistake was trying to take on too many projects - there was one group that wanted to grow the company quickly, because they felt that our financial problems would get better as we scaled up. Instead what happened was that talent got diluted - we kept hiring programmers, artists, writers and designers, and many were good, but many were not - including a number of "negative productivity" people.

But still, it was a good run - 5 years, and over 80 employees when it finally ended.

Very little.

I didn't know about the German version until after it came out on the market, but I do still have the German magazine featuring it on the cover.

Yes. For example, there was Land of Legends, which has a whole long history. My friend Joe Pearce had written a little dungeon game on the Apple II, and he called me up one day and asked "how can I get a job in the computer games industry?". I got him to drive up to LA from San Diego and meet with Jim. I had just finished FTA, and Jim was wildly enthusiastic about doing another fantasy-based RPG, this one using a first-person turn-based perspective similar to Bard's Tale. However, during the discussion Jim started brainstorming a huge shopping list ("I want to be able to build bridges over rivers in the dungeon, and I want to have things like pack mules that you can take with you") and us poor naive saps took in every word and promised we would build what we wanted. Two years later, I told Jim to cancel the project - it was simply too ambitious and we weren't ever going to be able to finish it. Instead, Joe and I would give him an updated version of Discovery, which we were able to do in a relatively short time frame.

Note that my involvement on the programming side was relatively small, as I was working on Music-X at the time. Joe moved in with us, so I was able to give him help from time to time. Ken Jordan also worked on the project, I don't remember who else. Note that Joe has had a long and illustrious career in the games industry since that time, he now sells copies of Inherit the Earth on his website at

No. We had worked tirelessly for two years but we kept cutting back the scope of the game due to financial pressures. By the time the game was near completion, the Dreamers Guild was going bankrupt, and I worked for three months without pay to finish the project and get it into a publishable state. But the game was a pale shadow of what we had intended, and both us and the publisher were in chapter 11.

At the same time, however, we had an incredible team of people and had developed what was some amazing technology for the time. Isometric tile systems with multiple cutaway "floors", 3D rendered character artwork that was then hand-painted over to give a more "painterly" look, dynamically-recolorable animated characters (so each Orc could look unique), and so on. There was a lot to be proud of.

This was the first time I had ever been in a leadership position with so many people - I was directing something like 30 creative programmers and artists I think, and I was dashing around the office so much that people would send messages to me by slapping post-it notes on my body as I rushed past them.

The failure of both the company and the game - and my breakup with my girlfriend, Allison - left me emotionally drained and burned out for many years. It was at this point that I started to succumb to clinical major depression, and it got worse over the next seven years until I finally got up the courage to go see a doctor and get on anti-depressants, which I still take today. In addition, I migrated from LA to the bay area, which meant leaving behind a lot of the friends I had made there. I've never made friends easily, and being depressed means that I wasn't willing to spend the energy to maintain my existing friendships. As a result, my "social network" gradually decayed to almost nothing, and I began to suffer from bouts of intense loneliness and isolation. Only when I was working on a technical problem - so absorbed that I didn't have any brain left over for sadness - was I able to function normally. (Which was fortunate for me, as it meant that I was able to remain employed.)

I've never really recovered from this. Although with the anti-depressants I am able to live a fairly content life most of the time, it's been something like seventeen years since I've been in a full-time relationship, and I still have occasional "episodes" of bleak despair. (Recently the Google+ team went on a trip to Maui and I had an emotional breakdown in the middle of it - I had to hide in my hotel room for hours at a time so that my co-workers wouldn't see me crying.) Fortunately, this only happens for a few hours every six months or so, so it's something I can live with.

So I guess the statement "I was not happy" would be somewhat of an understatement.

During that period I worked for a number of startups - Postlinear Entertainment, Transactor Networks, Brodia, Explorati, all of which failed or ended after a year or two - until I finally landed in a place of relative stability which was Maxis. I used to tell people that my life was like a platformer game where the platforms are collapsing under your feet as you jump on them - except in my case the platforms were companies.

Actually, I spent about a year working on Sims 2, and then another year working in EA's central technology group. It was the decision to disband the central technology group that finally got me fed up with EA so that I decided to leave.

At the time, I wasn't sure what I wanted to do - but I wrote an email to a number of my friends asking them for advice. Guido Van Rossum (the creator of Python) and Michael McNally (the brother of Robert who co-founded the Dreamers Guild with me) both suggested that I come interview at Google.

A few. Unfortunately, I don't own a television or a gaming console, otherwise I would like to try out Skyrim. My main desktop machine is a 27" iMac, which is great for doing development, but not so great for getting the latest and best games. I particularly like role-playing and character advancement games - I played World of Warcraft for about 5 years and I have something like 6 level 80 characters, but I finally quit playing about two years ago.

I find that I'm much less tolerant of games than I used to be. For example, I recently tried both Company of Heroes and Call of Duty: Modern Combat (bought via the Mac App store), and while as a game developer I am really impressed with both games from an artistic and technical standpoint, the fact is that as a player the games just bore me to death - I really have no interest in combat for combat's sake, I much prefer games where there's something I care about at stake, or there's some over-arching goal that I can work towards as opposed to just some sequence of scenarios.

I'm a software developer working on Google+, and I do a lot of work on the user interface. I was one of the people responsible for the recently added "nicknames and pseudonyms" feature of Google+.

I've seen the original FTA running on an emulator, and I was quite impressed. I don't have any Amigas - I'm the kind of person who is always looking for the next thing, the next project, rather than dwelling on the past. Although this also means that I'm not terribly good at self-promotion.

Actually, one of the things I am working on right now (as a side project at home) is a way to create a generic game engine using tiles, similar to Faery Tale 2 - except that the tiles themselves will be 3D objects, viewable from any angle. You see, the hardest part of making any game is the artwork - the characters and scenery. But what if you could have a standard set of scenery components that could be arranged in different map layouts by a non-artist? And what if you could have a set of stock animated characters which could be recolored and customized by a non-artist? That would allow anyone with a modest amount of programming skill to create all kinds of games - strategy / war games, adventure games, role-playing games, and so on.

So yes, it would be possible to create something like a Faery Tale 3 on top of that engine without too much difficulty.