Interview: Jamie Fristrom
by Lars Doucet, January 25th, 2008 (Edited by Derek Yu)
Howdy everybody! My name is Lars A. Doucet, indie game designer, and I’m here interviewing Jamie Fristrom of Torpex Games!
Some background about Jamie:
Jamie Fristrom is an old industry veteran. He’s worked on everything from DOS games on the PC back in the day to Spider-Man 2 on the Xbox. Currently he’s working on a game for the Xbox 360 called Schizoid, a quirky blend of Ikaruga and Pac-Man that’s being billed as “The Most co-op game ever.” Schizoid also has the honor of being the first game developed in Microsoft’s new XNA Game Studio to be commercially released on Xbox Live Arcade.
In addition to these exploits, Jamie Fristrom is perhaps best known for his “Manager in a Strange Land” column on Gamasutra, where he talks about everything from what kind of people to hire, how to scope a game, and everybody’s favorite most-dreaded topic, Scheduling. He’s worked on big games on big teams with big publishers, and now he’s working on a small, much more “indie” flavored title.
He’s worked on 3D games, 2D games, and the pixel art games that so many of us in the indie game community (myself included) emulate in our own games. There’s a lot to be learned from the other side of the game development community, and so I struck up a conversation with him and here are the results. So Jamie, here’s some questions for the interview:
LARS: Jamie, once upon a time you worked on DOS games like Magic Candle. In between now and then you’ve seen the transition to modern-day “next gen” systems and had credits on everything from the Spider-Man movie tie-in games to Tony Hawk’s Pro Skater to Psychonauts.
What was it like in the age of the DOS game? What sort of disciplines and tricks of that era are still useful to us today?
JAMIE: Still useful? Good question: even back then we had debuggers – and you’d have coders who refused to learn to use them, or refuse to take the time to get them working with their codebase. Debuggers, along with high-level languages, are silver bullets. They made us so very much more productive. I still sometimes see people who don’t use debuggers. Use them!
LARS: What sort of crap did you have to deal with back then that you don’t because of the advance in technology?
JAMIE: What really sucked was 16-bit addressing space. 16 bits can only address 64K of RAM, but our games were bigger than that, and the 8086 processors did this weird segmentation thing where one word would address a segment and another word would address the contents of the segment – so all your data and code had to be chunked into tiny blocks, and whenever you added two ints together you’d stop and check to make sure they didn’t overflow, and you’d walk backwards 20 miles in the snow.
LARS: What good lessons has the industry learned, in terms of management and design practices, since the DOS era?
JAMIE: Everything we know, seems like. I have a list here: http://www.gamedevblog.com
LARS: What lessons should we have learned but didn’t?
JAMIE: I wish I knew; I probably haven’t learned them either. If I had to guess, if there’s anything I know but isn’t standard practice in the industry yet:
- Waterfall & big-design-up-front is useless for creating games. Might be good for ports but that’s about it.
- We need to break down the silos of artists / designers / and coders and get small, cross-functional teams making tangible, holistic contributions to their games.
LARS: Unrelated to the above but important: what was your role on Psychonauts?
JAMIE: I did absolutely nothing for that game, though I wish I could say I did. One of their programmers called me up and said “Hey, we could use help on our animation compression.” I said, “I don’t know much, let me put you in touch with a guy who does.” They talked, didn’t end up using that guy’s technique, and gave me a credit anyway.
LARS: I first started following your writings with the “Manager in a strange land” column on Gamasutra.com.
When did you first get into management?
JAMIE: I used to get ‘lead programmer’ credits on games because I was the only programmer. Then I got the chance to make my own game because the company I worked for was really unfocused and ready to overextend itself – they were pretty much letting every coder at the company make his own game, and the art staff (of, like, five people) were this pool of resources we were supposed to somehow share. So this game was my first real taste of leadership and boy did I screw that up. I must have been 24 and had no management training and hadn’t read a single book on management. We floundered for several months as I invented technology and tools without ever managing to make something that could be called a game. Then the company went under, saving me from future embarrassment. Later I read some books on management and software engineering just for fun, really, and started instituting better processes at the companies I worked for, and eventually got my first real leadership position – lead programmer on Tony Hawk for the Dreamcast, and this time I was ready and it was pretty much a slam dunk.
LARS: The audience for this article is the Indie games community, in which most teams consist of a coder or two and the occasional artist or musician. Could you please explain exactly what a manager does and why this role is important?
JAMIE: With a small team – which is what we have for Schizoid right now – everybody has to wear multiple hats. So you might get by without having an official ‘manager’ per se, but there are certain things that need to be done, balls that must not be dropped:
- You need to schedule, so that you can recognize when you’re in trouble and take appropriate measures, whether it’s to slip or to cut or to find more guys to work on the project. I do this for Schizoid.
- QA. Programmers, designers, and artists are probably all too busy implementing to make sure the build works, so if you have a producer this is usually something that falls to them – whether they do it themselves or hire someone to do it. If everybody on the team is an implementor, then they have to consciously set aside time to do the QA. Bill does this for Schizoid.
- Interfacing with the outside world. At some point you’re going to want to sell your game (or enter it in a contest, or something) and this means talking to people. Once someone’s buying, whether it’s a web portal or xbox live arcade or a publisher, they’re going to have certain demands, and somebody’s going to talk to them for that, too. Bill’s our guy for this. (I’m a lousy salesman, and if it wasn’t for Bill Schizoid would just be a little prototype in a dark place on my hard-drive.)
- Facilitating decision-making. Not everyone on the team is going to agree all the time and you’ll need some way to resolve those differences. It shouldn’t be the manager’s job to lay down the law, but he needs to find a way to get decisions made.
LARS: You work on a small team now. What do you think small teams can learn from your experience in managing game development teams?
JAMIE: The same thing we did on Spider-Man 2 scales down to Schizoid – proof-of-concepts and prototypes first; a point where you (try to) stop changing game mechanics, and “go into production” – start making the levels that are going to ship with the game, probably throwing out the levels you’ve done so far.
I actually think the big team game developers have more to learn from the small teams than vice-versa, surprisingly enough. I have learned (and sometimes relearned things I’d forgotten) a ton of stuff working on Schizoid. The value of small, cross-functional teams; the fact that scheduling difficulty isn’t something that comes out of our large teams, but is simply endemic to game development, period; and I think Richard Garfield & Skaff Elias’s approach to level design could really enlighten some guys working on the twenty-million-dollar games. They could cover it a lot better than I could; I didn’t watch them do it; but some of the things they did included:- Just a couple of days of on-paper planning where they were just sketching levels. These were not cast-in-stone blueprints but back-of-the-napkin type stuff, and wasn’t so much to become the future levels but to “create a vocabulary of levels.”
- A large number of the levels are simply – “What happens when you bring two kinds of enemy together?” A level where you’re dealing with Flitts and Skulks is a totally different experience than Scorpios and Skulks. It sounds obvious, but I think back to Spider-Man 2 and typically, for “story” reasons, you would only encounter one kind of enemy at a time. (Except for the bonus combat arena you could unlock at the end – where you encountered everything at once.)
- Also, for a large number of levels they started with a title, and then figured out what the level was from that. “Let’s have a Spartans-defending-the-pass level, like 300.”
LARS: There’s a lot being said these days about Software management practices. Books like The Mythical Man Month, Peopleware and Joel on Software are frequently cited. However, Games are a unique animal.
To what degree are conventional software management theories applicable to managing a game development team? Specifically with regards to scheduling, waterfall vs. iterative development, and how you feel about all the new buzz-words like “scrum,” “agile,” and “test-driven development.”
JAMIE: It seems like the Cerny method came and went. Nobody talks about it anymore; everyone’s talking about Scrum now. The Cerny methodology is designed for what we do, and Scrum was designed mostly for in-house programmers creating non-shrink-wrap software, and is being adapted to what we do. Cerny and Scrum can probably work together, but we should adopt Cerny first, Scrum later. And – TDD is cool.
I use it and I’m not going back. I’d hesitate to call it even a copper bullet because it’s not an unambiguous, clear win, but I’m sold.
LARS: Could you explain why Cerny is better for what we do?
JAMIE: A big part of Cerny’s method is prototyping first, production later. Don’t spend a crapload of resources until you’ve found the fun. A lot of people think they can describe fun in a design document but you simply can’t, so waterfall doesn’t help here. And scrum leaves it up in the air what your priorities are – it’s up to the project owner – so a poor project owner armed with Scrum could easily say “We’ve got to get those displacement maps in there” when they haven’t found the fun yet.
LARS: In addition to “Manager in a strange land,” you also maintain a blog.
What have you learned by exposing your management publically and asking for feedback?
JAMIE: I seem to make a lot of bold general rule-of-thumb pronouncements and then, with the feedback, clarify and refine the ideas until they are subtle and nuanced. And sometimes I can go to the commenters just with questions and find good answers – which wiki to use; why C# behaves in a certain way; what’s the state of the industry relative to design docs.
LARS: Do you think that being a Gamasutra columnist helped Torpex games get more press? As in, Gamasutra sees the press release in a huge stack and goes, “Oh! Here’s Jamie’s new game!”
JAMIE: Torpex Games was getting press even before I was on board, possibly because Richard Garfield is involved, so I don’t think it was me.
LARS: I just realized you were talking about that Richard Garfield. I see what you mean!
You’ve worked on a lot of AAA titles over the years. Now you find yourself in a much smaller team, Torpex Games, working on a much more “Indie” title using Microsoft’s XNA Game Studio.
Some critics have said that XNA “isn’t ready for prime time,” doubting that it can deliver the “next-gen” experience. What do you think of XNA’s capabilities and how do you respond to these critics?
JAMIE: If “next-gen” means Halo 3 then I’ll agree, it’s not ready to make Halo 3. Is Geometry Wars next-gen?
LARS: Do you play indie games? Specifically, have you played Cave Story, Knytt, or Aquaria?
JAMIE: Not too much – I haven’t played any of those – the last ones I played were Baker Street and Flash Portal. (Mostly these days I play Rock Band and do cryptic crosswords.) My all-time indy favorites are DROD, Battle For Wesnoth, and Lugaru.
LARS: Can you speak a little about why you liked those games so much? What did they teach you?
JAMIE:
DROD – a new kind of puzzling. When encountering a new sort of puzzle, be it crate-stacking or sudoku or cryptic crosswords or minesweeper, the most fun part is when you’re learning the puzzle. Eventually you hit a point where you know all the rules and you’re just playing for speed, internalizing the rules, and that’s less interesting than just discovering the rules in the first place. DROD and Sudoku are good reminders that the sky’s the limit – there’s always new kinds of interesting mental challenges we can come up with.
Battle For Wesnoth – just really good, polished turn-based strategy with RPG elements in a fantasy setting. It’s amazing that it’s open source. I wouldn’t have believed that an open source team could do anything other than clones like FreeCiv until I saw this.
Lugaru – Ben Cousins once timed the jumps in a bunch of successful platformers and determined that they were all close to half a second. A lot of my friends and I really prefer games where the jumps are “big and delicious”, where jumping is almost flying: Galleon, Spider-Man 2, Hulk: Ultimate Destruction. This is another game in that camp – all these games show the value of sometimes breaking the rules.
LARS: What’s the appeal of the small team? What can it do that a big team can’t?
JAMIE: It’s more like a rock band. Every member gets to make a valuable contribution. It stops being “Joe Creative Director’s” game and you talk about who you have designing, doing the art, doing the music.
LARS: Let’s talk about that for a second – we’ve seen plenty of “Sid Meier’s” this and “John Carmack’s” that, and “Peter Molyneux’s” so-and-so and “American McGee’s” such-and-such. What are the effects of the “rock-star” marketing of games? Is it a good idea? Does it lead to strife on teams (“Oh, so Sid Meier gets all the credit?”)? Or solidarity and pride (“Dude, I’m totally working on a Sid Meier game”)? Is it a phenomenon you really only see in AAA titles? Anything else?
JAMIE: I’m torn, because I’d like to be a rock-star-of-gaming myself. (Or, better yet, just a rock star.) Partly why I blog – “Hey, look at me, I worked on this great game.” That’s how hierarchies happen, I suppose – the masses all think they can be president someday. I’ve seen both. Part of the reason I joined Torpex was to work with Richard, “I’m working on a Richard Garfield game.” That’s the best of both worlds because he’s an incredibly humble, agreeable guy – so we get all his genius but he doesn’t shoot down other people’s good ideas.
I’ve also seen people really pissed at their directors for acting like rock stars, though I should point out if you’re in a leadership position you’re going to piss off some people. An 100% approval rating doesn’t really exist. Politicians are happy if they get 55%, right? What I’d really like to see is rock band rather than rock star marketing, where the team is what gets the props.
LARS: What do you miss from being on a big team?
JAMIE: The programming manpower, kind of obvious, really. I’ve spent most of my time on this project programming stuff I would have delegated to a half-dozen guys on the Spider-Man team while I kicked back and drank gin and tonics.
LARS: In the Indie Games community, managed programming languages are quite prevalent. Game Maker, Multimedia Fusion, Flash, and now XNA Game Studio account for the majority of games we see online.
What made you go with XNA for Schizoid?
JAMIE: I hadn’t tried any of the others you listed above, so I don’t know about them, but I tried Pygame and Managed Direct X and Torque Builder and Panda 3D. Pygame was too slow and had weird bugs I couldn’t track down. MDX had a terrible api. Torque Builder meant I had to learn a new nonstandard langauge – I hear they’re C# friendly now. Panda 3D was slow, slow, slow. So XNAGS was the first thing that had both a beautiful language and excellent performance.
LARS: What advantages does a managed programming language have over traditional languages like C++, in your experience as a project manager?
JAMIE: More reliable builds; fast build times = fast turnaround; things that would crash in C++ just become nuisances in C#, so it’s great for prototyping. (Although, when you need to actually ship, you still need to fix a lot of those things.)
LARS: What are the disadvantages?
JAMIE: None of our coders had programmed in it before. We had a lot to learn. It’s not portable, though I do hear stories of people getting Mono to build and run on consoles. We’re dependent on Microsoft for a lot of stuff. We’re uncomfortably far from the metal.
LARS: Digital distribution seems to be all the rage these days, with channels like Steam, Xbox Live Arcade, and Nintendo’s WiiWare and Virtual Console seeming to take the industry by storm.
Do you think these channels will revolutionize the industry? Or is it just hype? What do they offer that traditional outlets didn’t?
JAMIE: Yes. It’s not hype. It was getting so games either had to be fifteen-million-dollar blockbusters or shoestring-budget indy efforts. There was no middle ground. Suddenly there’s an avenue for games that cost half-a-million or even a million dollars to be profitable. We can take risks again, invent new things. Who knows? Maybe we can bring back the adventure game, for example.
LARS: Many of these outlets promise to be the “YouTube” of games. What do you think of that metaphor? Is that a good thing?
JAMIE: I still actually want to make money with my games, but I’m glad that sort of outlet’s appearing, just for Flash Portal, for example. The economics are different, because making even a flash game is a hell of a lot more work than taping some video and downloading it…
LARS: Plenty of people talk about bringing down the “gatekeepers” who determine what games get published, and democratizing the games industry. Is this something we want to do? Is it even possible? (Obviously I feel we should do this, just want to get your feel for it).
JAMIE: Yes. Yes.
LARS: Let’s talk about Schizoid for a little bit:
Schizoid, in addition to being made in XNA Game Studio, looks from the trailers we’ve all seen to be entirely 2d. Is Schizoid entirely 2D? Or does it have some 3D stuff we’ve not seen?
JAMIE: Almost entirely. Some of the creatures explode into texture fragments that flip in 3d as they fly away.
LARS: Ever since Mario 64, the indie games community has been marching to the beat that “2D will rise again.” Did it ever die in the first place?
JAMIE: It might have. People look at you funny when you say you’re making a 2D game. It’s like you want to make movies in black & white or something. I myself was once on a “games are about immersing you in an experience” kick that I’ve recovered from. Now I just want to play good games that engage me, and 3D often gets in the way of that – most games are fundamentally 2D from a 3D perspective anyhow, and I wish they just made them 2D so I could see what the hell is going on. I like Phantom Hourglass better than Twilight Princess, for example.
LARS: Why did you make Schizoid 2D?
JAMIE: We’re all big on clarity. We want to know, almost down to the pixel, if our ship is going to collide with an enemy. 3D would make that murky. It’s the difference between Geometry Wars and Mutant Storm.
LARS: Phrases like “gameplay over graphics” often come from the indie gaming community. It’s a cliche I know, but what does it have to teach us? Are there any games out there that ignore graphics and sell on their gameplay alone? What else can you say about this age-old mantra?
JAMIE: The old Infocom text adventures used to sell on gameplay alone. (They would spend half their budget on marketing, interestingly.) I can say I agree with it – competing over best graphics is a bloody, red ocean.
If you have a unique gameplay hook you’ll be uncontested. One thing people fail to ask themselves often enough is whether their graphics are making their gameplay more clear. A game can be beautiful and have very simple, low-tech graphics…much like the graphs of Edward Tufte.
LARS: What do you think of XNA Game Studio’s graphical capabilities?
JAMIE: They’re good! Having HLSL underneath it all means you can do some real eye-searing stuff.
LARS: Will we be seeing Schizoid on any consoles other than Xbox 360? Or is it too early to say?
JAMIE: Too early to say.
LARS: Schizoid is being billed as the “most co-op game ever.” Co-op is personally one of my favorite kinds of game experience. Why do you think there’s been such a scarcity of good co-op games?
JAMIE: If your game is intended for multiplayer that means it’s probably not so good single-player, and games with a poor single-player experience have historically gotten crushed in the marketplace. We’re actually taking a pretty big risk with Schizoid, hoping that Xbox Live Arcade means that the old rules have changed.
LARS: My initial background was as a board game designer. I love board games because they bring people together; Some video games are good at this, like Smash Bros., Worms, and from the looks of it, Schizoid. Other games seem to pull people away from society, especially MMO’s. Video games are one of my fondest childhood experiences; at the same time, we see a lot in the media about how they supposedly develop anti-social and violent tendencies in impressionable young kids. And I definitely remember fighting with my brother over them.
Do you think we as game designers have an obligation to society for the games we make?
JAMIE: An obligation? No. But I think that game designers who care about the effects their games have on society are better people than ones who don’t…wow, that’s a non-answer, sorry.
LARS: If games are art, what does that mean about our moral responsibility for it? Does that absolve us of responsibility? Or does it heighten it?
JAMIE: It appears to absolve us of responsibility because you can do anything you want and then, after-the-fact, say, “That’s because it’s ART.” Some suspect the Super Columbine Massacre guy did just that – created something superficial and gratuitous and then later decided to pass it off as art. But I’d like to think it heightens responsibility, because people who really care, who play the games deeply, will notice the difference between, say, GTA3 and Manhunt. Though they’re similar on the surface, GTA3 is a work of art and Manhunt is snuff porn.
LARS: Please take this as an invitation to differentiate between what you think we “should” do, as opposed to what we “must” do. (Ie, what’s the right thing, on one hand, and on the other, what should we be held accountable to the government for, which are two very different things).
Beyond the classic, “do games make people violent” there are other questions to ponder.
- Are video games good for us?
- Are Skinnerian reward schedules (upon which the classic MMORPG “grind” is based off of) manipulating us and creating games that are little more than “Heroin Hero”, as seen in a recent South Park episode?
JAMIE: I do think we need to be honest about the negative aspects of games if we want to be respected for our stance on the positive. And I do think games are somewhat addictive – not physiologically, but the fact that we often feel the urge to sit and play a videogame, by ourselves, instead of working or spending time with our families is telling. But I don’t think it’s a serious addiction – I would put it somewhere around “caffeine” on the addictiveness scale, way below nicotine and heroin. That said, games are good for us. Almost all games teach us persistence, the value of practice, the nature of mastery.
LARS: And a few last questions:
At XNA Gamefest, you talked about how Microsoft really gave you a hand, even letting you work on a “double-secret-probation” version of XNA Game Studio with special features so you could speed Schizoid’s production along. What do you think prompted this on their part? (If you can talk about it).
JAMIE: They want people using XNA Game Studio. Why, it’s hard to say, because there doesn’t seem to be a revenue model for it, and I can only engage in idle speculation, but I think they’re trying to bring more people to C#. (And once you’ve used C# you don’t really ever want to use anything else.) And if they can say, “Look, these guys are pros, and they launched a game using XNA Game Studio, and they made money,” that’s one more reason for people to use it.
LARS: Did you play any of the finalist’s games at Gamefest? If yes, which was your favorite? And no, you don’t have to say “Sprockets of Strife” :)
JAMIE: I did, but I only played a few of them, and I didn’t play any of them deeply enough to responsibly pick a favorite.
LARS: Jamie, there’s some awesome stuff in here. Thanks so much!
JAMIE: Thanks for killing an hour! I knew I shouldn’t have checked my e-mail during that build…
And good luck with Sprockets.
LARS: Thanks!
