In-Depth: Far Cry 2's Guay Talks Dunia Engine, State Of PC

Author: by Christian Nutt, Staff

Far Cry 2, the ambitious Ubisoft Montreal-developed open-world sequel to Crytek's original PC game, was announced as a PC exclusive before eventually gaining Xbox 360 and PS3 versions as well. It is now expected to ship for all three systems this fall. Along with the game, the Montreal team is developing Ubisoft's new Dunia engine, multiplatform technology that the publisher plans to use in several of its studios for future projects. Dominic Guay is Far Cry 2's engineering director, and he spoke with Gamasutra at a recent Nvidia-hosted event about many aspects of the game's development. These include multiplatform issues as PCs being to technologically outpace consoles again, why the game's open-ended nature made it perfect for engine co-development, how Ubisoft manages the multitude of teams at the 1,500-person Montreal studio, and more. Team Management At the peak of this project, you had 150 people working on it? DG: Yes, that's right. Were those all people at Montreal? DG: Yeah. Did you outsource anyone? DG: No. We didn't outsource anything. I guess when you have a studio that large, that gives you a lot of freedom to keep it in the family. DG: Yeah, I think that we're really lucky to have a this size, but also this amount of projects. There are always projects and conception and production and preproduction - multiple ones. There are never layoffs because of, "Okay, I'm on a different project away from other people." That doesn't happen in the studio. There's a lot of job security for people on very different projects. This person can do that kind of game. So we're lucky. I realize that. How do you organize your teams? There's a lot of talk now about whether to have interdisciplinary structure on an office level, or if you should have people working together of the same disciplines and then try to have communication across. DG: Yeah, that's a pretty good question, especially since we have a very large team. If you have a very small team, you can do all of those things at the same time, because you're all sitting at talking distance and there's no problem. We have close to 150 people in production, and once you reach 100 people, you need to think differently for your organization. At Ubisoft Montreal, different teams tried different things. So we had a lot of flexibility. Initially, we were really structured according to discipline, and one of the things that we started doing is what we call "owner groups." For example, in our systems, if we want to make our driving kickass, there will be driving owner groups. There will be maybe one or two programmers, one designer, one animator, and one artist. They will work together, and there will be a leader to that group, and they will have a common goal that really brings them closer together. It doesn't mean that they don't talk to their people and discipline, but for a little while there, they feel more tightly incorporating with those other guys and girls than with their own discipline. So we have the owner groups, and we still have people working with their discipline and in groups. Do they move together physically when they enter groups? DG: Sometimes we do that. It's not always necessary. They can have daily or bi-daily meetings together and stand-up meetings and stuff like that, but when we can afford to do it, sometimes we actually want to group those people together. When you say stand-up meetings, are you developing under Scrum for this game? DG: Not in a pure fashion. Not according to the strict rules of Scrum. That seems to be pretty prevalent, too - people adopting it, but not totally. DG: Yeah, totally. There are a lot of things that are key to Scrums and key to Agile that we do like. We try to have a relatively short iteration. We define what we want to get done in a sprint or an iteration, and we always at the iteration have a version, and we can look at it together and get to say what they want and prioritize. There's a logic there, and we try to do those standup meetings, but we don't do it in the pure sense. It's adapted. Cross-Team Sharing Do you have a best practices group at Ubisoft? Do you share them? Ubisoft Montreal at this point must be a pretty massive collection of developers, with the number of projects you guys have been cranking out, and high-profile projects as well. DG: Well, there are around 1,500 people working there, including all the HR. Actually, a month ago, we did an off-site with all the leads. I think it was 60 or 70 or 80 people. We went in the woods in cabins, and we spent two or three days exchanging ideas on how we organize. Some people would say, "Okay, how would you improve or deal with this?" and other people on the team could say, "Well, that's the way we do it, and we deal with that program." That's the way we're organized, and we exchange ideas between leads. That was a pretty good activity, working out what we want to do. We also have people who are responsible per-discipline to try and improve the processes for them as well. It's a tough job to do, because everybody's head is in the project, and you just want to get it done, but most people are like, "Okay, those guys are doing something right, and I should tell the other teams that this is the way to do things. And maybe we should have one person who's free to go from team-to-team to propagate that good idea." We're always trying to improve that knowledge and technology sharing, but it's hard. It's super-hard to do. I'm not saying that we're perfect at it, but at least we try. (laughs) I think it's a first step. It must be absolutely difficult to find the time to have these things. If you guys are nearing the end of your cycle, but a team working on, say, Assassin's Creed 2 - which I'm just presuming exists - is in a very different part of their development cycle, it could be very difficult. DG: Yeah, it is hard. I went to that activity, and I was struggling with, "Should I go? Because I'm really busy here. I should go. It happens that this time is not the best for me, but next time it will be for someone else, and I need to take that time." And I went, and it was a good thing I went. It's interesting, because there are still a lot of models for studios. You work at what would probably be one of the largest studios in the world, I would imagine, with 1,500 people. DG: The second I think, after EA. Right. Then there are still small development teams that are independent, or medium-sized independent developers, or developers who have a lot of studios that they own, but they're in different physical locations. Do you have any insight into what works? DG: (laughs) I've visited a few developers since I'm from Ubisoft. I've visited some third parties who want to be published by Ubisoft, and I've seen all sorts of models. But I think that problems are common, so I think everybody's trying to find solutions. Everybody finds good solutions, unless there's universal good solutions. I think what's important is that these teams try to find solutions and try to organize. I think we're eventually all going to get them right. And this event you had, was that just for Ubisoft Montreal people, or was it cross-Ubisoft people? DG: That one specifically was just Ubisoft Montreal. Sometimes we also have a knowledge-sharing event at the cross-studio level, if you want, and we also have people at the headquarters who are responsible to do knowledge sharing. That's their job. But it's even harder for them, because then they have people in two different time zones who maybe don't speak the same language. I'm trying to think of studios that Ubisoft owns, and they have a studio in Japan that they bought a while ago. They're making DS tech games. DG: Yeah, and we have two studios in China. Right. And you have Romania that did Blazing Angels. DG: Yeah, and France, to the States, to Canada...yeah, a lot of studios. The things that Ubisoft is trying to do is instead of pushing stuff, it lets us hone. One of the things that they're doing at headquarter level is that they have this website where the people in charge of sharing and comparing information from every project and processes and tools can go there and quickly get references - "Okay, this guy has already done that." Looking at it from Montreal studio, we have a lot of room to improve just at the local level, you know. We know it, and we want to. So we start by focusing on that, but we still work with the other studio, because there are smaller studios of like 60 people, so they shouldn't be cut off. One of the things that Ubisoft also does is something called the Academies. They bring experts from all the studios, get them together outside, and they bring a subject and say, "Find a solution," and try to get them to think together. That usually is surprising. We realize, "Okay, we're all facing the same problems." We have different ideas, and we exchange about it. It's pretty cool. Fostering Studio Culture It must be a giant task to manage all that. DG: Well, the vision of the studio, to simplify it, is that because we want projects to innovate and be creative, there's a lot of flexibility. They put measures in place to favor creativity and innovation, and they don't guard them. They give flexibility to work things the way they want, and the studio offers services, basically. It feels, to me, like we have our small independent developer with all those facilities. I'm not saying we aren't spoiled. We are spoiled. But our own little independent developers have access to all those great resources around us, but we're independent in that we have our own strategy and our own way of working. That's pretty cool. This is even being reflected by EA now. They're saying that they've made mistakes in the past by trying to message outward with the corporate culture out to the developers that they acquire, and it stifles and in some cases destroys them. This is a fundamentally creative process, and you need to have a stimulating environment for people to create in, otherwise it's going to become very rote. DG: I think you're absolutely right. In the last year, I've been reading things from EA that I've been hearing for six years that I've always been hearing at Ubisoft. I think maybe they might have picked up on some of that. Engine Development This is the first game that Ubisoft is shipping with this new Dunia engine, right? DG: Yep. Was this engine developed in parallel with your game? Did your team develop it, or did you have a core tech team working on it? DG: We kind of have a hybrid of that. The engine team was embedded in the project. It's really embedded, so it's not like they're on a different floor and you can only talk to them by e-mail. I'm responsible for the engine that brought them, and I also have a key role in the game that brought them, so that kind of proves how much that was embedded. It worked pretty well, because that way, you can be sure that the technology that is being developed in the tools matches really well with the real production and the real meat of what's being made. If you have a distance, there's a risk that you'll have discrepancies between the needs and what's actually being done. And that worked out pretty well. But what will make this a success, in terms of engine development, is that the guys who are working on the engine are also very aware that after the game ships, it's not a done deal. That engine is going to be used by others, so you don't do something that only works part time. It's something you can reuse, expand, and improve. Basically, what we did is that since we had a lot of time, we hand-picked people to be sure that they're the best guys who can do it. That's always a concern with any engine, even ones that you license, that they're either keyed to a genre, or they don't have features that will allow you to go in and make a game. You have to build up those features. When you're working on engine development parallel to the game, how do you ensure that's something that will be worthwhile? DG: One of the things is, if you're going to do a pinball game and develop an engine at the same time, you're probably going to end up in the situation you're describing, where the other team that wants to do any other thing than a pinball game will [have problems]. But because our game is so expansive - it's open-world, steaming at continuous speed, you see very far and very close detail, and a lot of physics - there are so many things in there, it already just for the game has a lot of flexibility. If you look at all the other game genres, you say, "Okay, that's got to be included in this one." Of course, if you're going to make a racing game in the engine, there's stuff that you want to do in optimizations that are specific to racing games. But we can clearly do a racetrack! (laughs) And there's already a vehicle. DG: Yeah. You could prototype your game pretty quickly and then do those optimizations. Is that a conscious decision, to take a game that had a broad stroke of different design ideas and then do development against that, or was it luck that you're making this ambitious game and that was the engine you used? DG: Well, the mandate was given down to me by headquarters. I can't pretend that I'm in their head and I know what they're thinking, but I think that when you're given a mandate to develop a technology, you probably take into consideration, "Do they have the skill set to do it? Do you trust them to be able to do it? Is the game they're making significant for the kinds of features that we'll put in?" They knew we had pitched to them the kind of game that we wanted to do, so they'll probably conclude, "Okay, this game is ambitious enough that it will support an engine that we can re-use." Ubisoft tries to do a lot of games that give a lot of freedom to the player. That entails an open-world kind of logic. So I think that wasn't a bad decision. High-End PCs And Consoles This is a high-end PC game, and that's an interesting place to be right now, but it's not PC exclusive. Dominic Guay: No. We kind of cover all our bases. Obviously, our key focus is going to be on the PC, and it's pretty scalable. We haven't defined any minimum requirements yet, but we have the game running on a Pentium 4 and an [Nvidia GeForce] 6600. Obviously, it doesn't have all the bells and whistles of what you just saw there, but the key experience is still there. You have the same game on that system. But we'll also be having a PS3 and 360 version shipping at the same time. For the gamer that has a maybe four- or five-year-old PC, he can play it on his PS3 or his 360. Do you think that it would be dangerous if it were only a high-end PC game? DG: If it were only a high-end PC game? Considering the amount of time we spent on that game and the investment, yeah, it's probably not a good idea to make an only high-end PC game, because you'll only limit your market a lot. That seemed to be something was more feasible in the past. The market for high-end PC games is kind of shrinking, piracy is big, consoles have become more capable. DG: Well, yeah, but games were always scalable on the PC, to a certain extent. Games usually let you have a two- or three-year-old PC and still be able to play, even if it isn't optimal or anything. There have been examples of, at the time, high-end games that required almost for you to have a brand-new PC. That might not be possible anymore, actually. Not in this period, for sure. They ran into some problems last year with Crysis. Crytek did come out and say that they don't believe that they can stick to being PC-exclusive for their games in the future. Obviously, you've already made the decision not to. DG: Yeah, I've read that. I think they referred to piracy as a major issue for them. Their game was perceived as being very high-end - and it is true that in order to get the kind of quality we were used to seeing in their videos, it did require a pretty high-end PC - but you could play their game on a lower-end PC, and you could still get a very good framerate. I don't know how much of it is piracy or not. It could be audience perception as well. DG: It could be perception. It's really hard to say. We're not going to take a chance here, so we're offering it to everybody. I think that someone who has a high-end PC and a console will probably prefer to get it on the PC, if he likes shooters. But people who have an older PC might prefer to get it on console, to get a stable experience. It's the same game, and it'll have solid 30 frames per second everywhere. When the current console generation started, back when the Xbox 360 came out it was on par with a high-end PC. This is something we've seen time and time again with console generations - as it moves forward, the PC starts to outpace. The new GPU generation is significantly improved as well. Is that a struggle as a developer making a game that's not shipping until later this year? DG: Yeah, we're shipping in the fall. For our timeframe, it works out pretty well. It's true that a card like a [GeForce] GTX 280 has a huge amount of processing power. But if you look at the actual install base, right now the install base on the PC can target something that's compared to console. We put a lot of investment and time in putting in the higher-end features that we can run on those higher-end boards, but for someone who doesn't do that, you can still target PC and console. Maybe in three or four years, that's going to be harder, because your game on PC will look like, "Whoa." I think that Nvidia is trying to help developers do those extra things on PC, and offering more graphically that doesn't stop developers from having the same game on consoles. Some of the gameplay you showed was based around the way the fire was directed by the wind. You can't not have that in any version of the game, right? DG: Yeah, it's there on all versions. Where we'll play is on the amount of detail in the visual ambition of that. But we won't remove any of those systems. You're trying to come up with gameplay systems that are really compelling to the broad audience of everyone who can play your game, but then you have almost a bonus for people who have the high-end cards. How do you weigh those options? DG: Like I said, for us, the timing made it possible for us to not have hard decisions to take. I think that, for example, for the next game, if we ship and we want to have a consistent game across all platforms and the PC like we're doing now, it might be harder, because in general, the PCs will be much more powerful. The good thing is that the kind of possibilities in terms of gameplay that we're opening up on the 360 and the PS3 are pretty wide, so there's a lot of unexplored area. It's not like, "Okay, we maxed it out." We're far from having maxed it out. So I think we can for a few years do new cool things, and have room to improve within what those consoles can do, without having to say, "Okay, we've done it all now."