Rich is the author of quite a few bots and mods for various FPS games.
celtic88: You have accomplished much in the field of AI and gaming, where does
all this wisdom come from?
Rich: I suppose I've been playing games since I was 5 or so, and I got into PC gaming when I was around 7 or 8 (if I remember correctly). After that, I got into programming very soon (I believe I was only 10 or so), and I started learning C/C++ when I was around 12 or 13. So, I'd say all my "wisdom" just comes from the years of experience. Between age 13 and now (18), I've focussed greatly on AI and have worked on a good number of bots. Each time I make a new bot or go back and work on an old one, I try to take what I've learned so far, then come up with a new idea and figure out a way to do it. Often, the specific methods I come up with for making the bot do something realistic can be carried over between games. A good example is how the rocket jumping priority code from my Quake bot (note that this code is actually not in the current released version of Tsumi to Botsu) is completely interchangeable with the gauss jumping code from my Half-Life bot. In the years I've been working on bots, I've come up with so many "interchangeable" ideas that I can practically make an advanced bot for any game almost only by copying and pasting code from my other bots. Although, I do often prefer to just write everything from scratch to avoid possible bugs or unoptimized things I've overlooked from previous bots (which is why I ended up writing the hidebot code from scratch).
celtic88: How does it feel being the author of the first ever quake 2 bot?
Rich: I guess it was a pretty good start for me, as far as publicity goes. The Famkebot was pretty sad back when I first released it and wasn't much more than a monster that looked somewhat like a player. I learned a good amount through the process of its development, though, and the final product ended up being acceptable. Compared to some of my more recent bots, though, Famkebot is pretty lacking in a number of areas.
celtic88: Do you still get feedback on the famkebot?
Rich: Very, very rarely. Every once in a while I got an e-mail from someone asking me how to start Q2 with the Famkebot or where to get route files. Sadly, there is no place to get Famkebot route files (that I know of).
celtic88: Will you ever go back and fix or add some things that may have been brought up?
Rich: With the Famkebot? Perhaps some day when I've got nothing else to do if it sparks my interest. I've got a good amount of other stuff to work on right now, though.
celtic88: What are some of the things you've learned from working with the famkebot?
Rich: I developed a very primitive form of the navigation system I still use in the Famkebot. That's probably the most significant thing I got from it. Other than that, though, I got a lot of experience with AI programming and programming in general.
celtic88: Are you still working on the Jumbot?
Rich: Not at the moment, though it's not completely impossible that I could go back and do some more stuff to it (it is in rather bad need of some speed optimizations).
celtic88: What kind of enhancements can we expect for later versions?
Rich: Speed optimizations.
celtic88: How long has it been since development initially began?
Rich: For the Jumbot? I started working on a sort of "generic" AI base that I planned to use in it a few months (again, if I remember correctly) before the SDK was even released. So, it's been development since whenever the SDK was first released and then a few months before that.
celtic88: The addition of bot leaders was impressive, any plans to add them to any of your other bots?
Rich: Possibly, since the code is all rather simple. If I work on another bot for a game that's based a lot around teamplay, I might very well use some of the bot leader code from the Jumbot.
celtic88: Perhaps the most realistic of the features is bots joining and leaving your server at will, any plans to improve upon this or add it to another project?
Rich: If I work on the Jumbot more, I might think of a way to improve on it. As far as using the idea in other projects, it's perfectly possible, but that's the sort of thing I don't add to a project until the project is virtually finished and all the AI stuff is pretty much complete.
celtic88: Has most of the feedback you've recieved been positive?
Rich: It varies from project to project, and depends on how complete the project is at the moment. I don't think I've gotten anything but positive feedback from later versions of the Jumbot, and no one seems to have any complaints about my unofficial hidebot release. I used to get a lot of complaints about earlier Jumbot versions, though (rightly so, anyway, since they did have a number of problems), and people weren't too happy when the hidebot that came with SoF Gold didn't support all the play modes.
celtic88: What's the craziest suggestion you've ever got?
Rich: I've gotten a terrible number of them that are just sort of silly, and people often tell me to basically write an entire mod on top of the Jumbot (which I don't want to do, since the Jumbot is just a Half-Life DM/teamplay/coop bot). The first odd suggestion that comes to mind is the suggestion to make suicide bots that would run at you screaming and then explode, though I sort of liked that idea myself. I'm sure I've gotten suggestions that were "crazier" than that, but I'm unable to recall well at the moment.
celtic88: The Scientist Hunt is another great mod, is it still in development?
Rich: Yeah. Version 1.1 will have a terrible amount of bug fixes and new stuff, and we've already got a good amount of it done (including the handcannon). I'm currently waiting on the chainsaw model to be finished, and then I'll be waiting on the scientist limb weapons to be completed.
celtic88: How did you come up with the idea for it?
Rich: Hondo and I wanted to do another violent mod together. We had no one to model for us and had to come up with an idea that used original game material (as usual), so we decided we'd make a mod where two teams had to track down scientists through the level and slaughter them in terrible ways. We got the basic structure of the mod done and released it, and we've been building on it by adding weapons and play modes ever since then.
celtic88: Which of your projects consumes most of your spare time?
Rich: At the moment, I've just been playing through Final Fantasy IX while I wait for the Scientist Hunt models to get done, so I guess nothing is taking up my time at the moment. The next project I start working on again, though, will almost certainly be Scientist Hunt.
celtic88: You have delivered joy to thousands of 56kers with the only Soldier of Fortune bot, any words for your fans?
Rich: Thank you for putting up with the first release.
celtic88: Is the Hidebot the most advanced bot you have worked on?
Rich: I'm not sure if I'd say that or not. It does have a navigation system that's somewhat better than the navigation in my other bots, but the Jumbot has a few more general "features". I guess the hidebot and Jumbot are both slightly more advanced in their own different areas.
celtic88: What are some of it's features that seperate it from the other bots?
Rich: The hidebot's main different feature is the optimized navigation system. Most if not all of the speed optimization ideas came directly from Keith Fuller at Raven.
celtic88: Some time ago you were hired by Crytek studios, are you still with them?
Rich: No, I quit quite a long time ago.
celtic88: Can you give us any information on what you are working on?
Rich: I guess I already said a lot about Scientist Hunt, but other than that I'm essentially "between projects". I guess the next thing I work on will be an Action Half-Life bot, but I'm waiting for AHL B4 before I start on that.
celtic88: Will AI take over the world?
Rich: Probably.
celtic88: When?
Rich: I guess it depends on how long it takes the required hardware for building bloodthirsty robots that can navigate reliably through areas and detect humans to go down in price.
celtic88: Will you be one of the mad scientists behind it?
Rich: I do hope so, but I don't think Cassie will let me.
celtic88: Some may say "I can't believe it's not human." when playing your bots, what's the secret?
Rich: I guess that would also be due to the constant improvements in general over the years. It's sort of hard to make a bot both challenging and human-like at the same time, but I think I come closer to the goal with each new bot project.
celtic88: Any last words?
Rich: Cassie always seems to be slightly upset when I never mention her in interviews. So, I love you, darling.
