An interview with independent game developer Sean Barrett.
A veteran of PC game-maker Looking Glass Technologies and their titles Thief, Terra Nova, and System Shock. Sean here muses on his past work with graphics and his interest in "interpersonal physics" - how games might better simulate engaging player-to-non-player-character relationships.
At LGS, I worked as a technology programmer, developing most of the graphics engine for Thief and adding significant enhancements to Terra Nova. I was not a designer on either project. Most of the programmers at Looking Glass Studios had design skills, but they didn't build levels or perform typical design tasks; however, because they had control over the code implementing a system, they would often do design work for that system--e.g. the person programming the combat system would figure out exactly how the combat system should work. Because the graphics work I did didn't have any gameplay elements, I came away from LGS (my only industry job) with a lot of design knowledge from working with people, but not a lot of design experience.
A fair amount of what I've done since then is give myself experience at actually doing design; for example, I released a small puzzle game (Chromatron), from which I really internalized a lot of lessons about player experience, ramping up difficulty, and such.
It sounds as though your work thus far in the games industry has been preparatory: learning, exploring.
I guess this is sort of a misrepresentation due to hindsight. At the time when I was working at Looking Glass, I would have said that my work was being an expert graphics programmer and accomplishing tasks that very few other people could. Coincidentally, at the same time, I had my hands in the design pot a little bit, just like pretty much everyone else did, so I did learn a bunch about game design then. (Although I also learned a bunch about game design theory when from running a MUD and even from writing some simple games in BASIC in high school.)
Over the years, though, I started feeling like competing on graphics--the stuff I was particularly good at--was the wrong way to go. In truth, I never really thought of it that way at LG--I knew LG was trying to push game design, and I saw the graphics stuff as simply enabling the LG design teams to accomplish the interesting design goals. But it felt to me like the industry really was starting to compete on graphics, which was particularly odd as graphics started converging, with graphics hardware, to everyone doing pretty much the same old thing. And I started to feel like, there was no point in trying to compete graphically with the people doing simple games and high-end graphics.
It took me several tries to break loose from the industry, and now that I am independent trying to do both programming and design on my own I'm well aware of how my design knowledge was improving while I was at LG, and it makes the "working in the industry" part certainly seem sort of like practice for "working independently", but I certainly never think of it that way. Working independently has been born of frustration with trying to work through the industry.
You've submitted a game design to the Indie Game Fest. And you made Very Serious RoboDOOM at the last Indie Game Jam. Is there a common theme between them? Homage to games/game types you've appreciated in the past?
Right now, I'm definitely caught in a trap of re-implementing game designs that already exist; whether that's all I'm good at, or just because I don't have enough experience yet, or I'm just being lazy, I don't know. Even before I went indie, this was a little true--I implemented several "minigames" in System Shock, all of which were reimplementations of existing games. RoboDOOM was an extreme: a homage and yet a kiss-off to the mindless shooter genre--I've certainly gotten many hours of enjoyment from that genre, but there's such a thing as going too far, or really, maybe, not going far enough. I currently have two games in development, and they're both strongly derivative, at least at the core, of two of my favorite games, the X-Wing/Tie Fighter games and the Ultima Underworld games.
I don't perceive of this as homage so much as trying to evolve gameplay a little at a time due to my limited skills, but also working with a style of gameplay that I know will be familiar to an audience already, so I have an easier time selling it to them. Once I have an audience it'll be easier to deviate (i.e. in sequels).
As you're putting the finishing touches on your current game (maybe?) what might be next? What game idea is percolating in your head? What technology or gameplay idea would you want to experiment with?
I'd like to find ways to revamp gameplay significantly, but that will get easier when I have stable technology I can build on for future games. For example, in my current game, I'm trying to implement a very limited version of this faction simulation idea Doug Church (also from Looking Glass, e.g. the project leader on System Shock) has been talking about for a while. I'm keeping my ambition low because I don't want to take huge risks or invest a lot of time, but it's been so poorly explored so far that I think any attempt at all will get us (or me, at last) some useful information about it. If it comes out interesting then I may try to push harder on that in future games. I'd like to push on character interaction--even if a conversation engine still consists of a message from the other character and a choice of responses for the payer, running this with some sort of character "simulator" which actually has some meaningful state and attitude and reacts to the conversation, rather than just being a menu maze that you navigate no differently than a voice mail system. There's been some interesting stuff coming out of the hobbyist text adventure community on this front that I'd like to follow up on. But really this is a technology issue as well as a gameplay issue--we don't really know how to make this sort of technology, and we need people to figure out how to evolve it. There are numerous other possibilities from a gameplay front; I'm a musician, so I'd be interested in a beat-matching or music-creating game of some kind.
I don't usually tend to work from a "ok, I've come up with a neat technology that lets me do X, so now let me find a game to do with it". In fact I have a history of coming up with neat graphics engines and doing nothing with them. The one technology that's on my radar right now is that Douglas Hofstadter (author of Godel, Escher, Bach) is an AI researcher who published a book, based on research he and his grad students did, about how to have a computer be *creative*--the book is called "Fluid Concepts and Creative Analogies". The technique is fairly interesting, so I have my eyes open for opportunities to try to use it. The most obvious scenarios--a place where we need a computer to be creative--are in "dungeon mastering AI"--where we want the computer to take a more active role in moderating the player experience, instead of just letting the player move through a world laid out by a designer trying to anticipate and code for everything a player might do--and in character simulation, since people are creative. But I don't see much opportunity in the latter case, I guess because people aren't creative all the time, or at least the sorts of situations where we normally think of encountering NPCs--e.g. asking a peasant which way Black Francis went. And the former is a big huge can of worms.
Can you look back over the technologies you've studied and the problems you solved and generalize about the specific themes that interest you? With regards to genre, perhaps, or programming issues, but more specifically, game play and game mechanics.
I don't think I've done enough serious work on the game mechanics/game play front to be able to pull anything out, beyond the issue you raised below: I definitely believe in strongly in simulation. For example, Terra Nova was this squad combat game I worked on before Thief, and I tried to get the AI programmer to make sure the enemies wouldn't cheat about knowing where you were, so you could potentially sneak around, lure them away and sneak back, etc. It wasn't a focus of the gameplay, though, so I think they did end up cheating.
Much of your current curiousity seems to lie in developing more rich simulations for participants of single-player games. Certainly I found the characters in Might and Magic VI to be wildly bogus and silly. What did I expect? At least Baldur's Gate has some massive scripting and some cute touches. But as you observe, there would seem to be some rich potential for games to better engage us with state-based characters and more customized environments. Do your two current games play with these ideas of enhanced game responses?
Not very much. The RPG is a lot of work just doing "old school" conversations; my hope is that if it were to do well, then improved characters would be the top priority for the sequel--I wouldn't have to write a graphics engine or a lot of other things, so there'd be room to do a bit of research. The hovercraft game I'm trying to add a little bit: the idea that your squadmates have personal agendas (derived from their allegiances to various political and religious groups) and become satisfied or unsatisfied with you based on whether you cooperate with those agendas *in game*; so one squadmate may desperately want to know what's in a bunch of crates, and another may want you to blow them up without finding out at all, so you can blow them up, or identify them, or identify them and blow them up, or just ignore them because it's a total distraction from the main focus of the mission. (This is the "faction" stuff that I mentioned before.)
So the approach here is pretty simple: each participant just either likes or dislikes you in varying degrees, and has certain canned attitudes to take depending on that degree, as well as a number of tasks to assign you. Rather than a branching conversation, your interactions with them are purely through actions you take in the world. That gets rid of the distracting and limited conversation menus--replacing them with an even *more* limited set of inputs specific to that character, but ones that are a subset of a totally free, continuous set of arbitrary actions you can take in the game world.
I definitely think it's an important area to explore, but it's also not a panacea. If we start having NPCs who have goals and wills of their own, that could easily get in the way of having *fun-for-the-player* gameplay. It wouldn't be that fun in an RPG to have a "really conversant" NPC whom you have to chat with for 45 minutes, making friends and building trust, until they'll give you the key you need--unless that 45 minutes of conversation is totally fun. But even then, if that conversation feels like "I am playing a game to try to acquire a key" not "I am trying to manipulate this person into giving me the key" it would probably be silly no matter how fun it is, and not really fruitful. Where I think this NPC stuff is important is in the genres where videogames, as a dramatic medium, aren't keeping up with any of the other dramatic media--we do violent conflict great, but we don't do any other kind of conflict--especially interpersonal relationship sorts of conflict--at all. Because we know how to do violent conflict so well, it makes sense to me to try to integrate the two to start with--the violent conflict is the plot, and the interpersonal stuff is in subplots--but ideally we'll start finding we can create games that aren't about shooting things. (I mean, as clever as Mario Sunshine and Pokeman Snap are for subverting what it means to "shoot things"--making them non-violent--they're still about shooting things.)
(This is an exageration: there are non-violent conflicts. Sims for non-violent sports like golf, racing, etc. -- but here the conflict isn't interpersonal, but rather conflict owing to a fixed set of invented rules in the real world -- we are just simulating real-world games; abstract puzzle games have no real conflict. Adventure games, both graphic and text, are probably the strongest contenders here, but the characters feel no deeper than Baldur's Gate and what not -- the characters are essentially non-interactive, the interesting actions being entire scripted.)