Interview on HalfeR.net

The russian Half-Life fansite HalfeR.net recently invited us for an interview. They had  been following our project straight from the start and were curious to find out more about our team’s history, work structures and plans for the upcoming release. In addition, you can find three brand-new exclusive action shots of TNE on their site.

Check their article for the extensive interview and images.
If you can’t read cyrillic, we have an english transcript for you.

Actually, it is not – the inspiration for the project came from an entirely different title, which dates back to 2003. Jedi Knight: Jedi Academy was a very popular title back then with a very vivid online community. Lindsey and I (the founders of TNE) were level designers for the Moviebattles II mod back then, which until today is one of the longest-lasting and successful mod projects I’ve ever seen. Moviebattles always aimed at being the most competitive Star Wars online game around with objective-driven gameplay, lots of customization and a focus on teamplay.

At some point however, the id tech 3 engine became rather aged and Lindsey and I could no longer realize our visions. Back at the time, we tried to create a prototype for a game mode which today, you’ll probably know as “payload” in Team Fortress 2. That was before TF2 was even announced, but the limitations of the engine forced us to drop the idea.

The actual kickstart for TNE was in October 2005 when Valve published the “Lost Coast” tech demo for Half-Life 2. We were thrilled by the visual possibilities and the idea that you no longer needed a particular game to run the mod, but a Steam account with Source SDK base instead. We saw the opportunity to draw from engine updates without forcing our user base to move to new base games. Thus, we decided to give this a try and started developing in 2006. At first, the project carried the name “Moviebattles: Source”, since our initial goal was to create a successor of the mod on a new technology base. We ran into creative differences with some of the original MB2 developers and decided to rename our project, which would also grant us more creative freedom to do things differently.

Currently, we have 12 active developers. But as with all noncommercial projects that last so long, we’ve undergone so many iterations over the years that people come and go. In total, we’ve had about 60 people contributing to the project. That being said, most of the contributors are artists, which always leaves us short on coders. Thus, we don’t really have the capacities to assign people to different game modes. Our current coder Øystein Dale has rewritten the entire gametype framework in order to allow level designers to influence a lot of the game logic. While that takes away pressure from the coder, it also allows us to work out a wide variety of gameplay scenarios without having to worry about technological feature creep.

We had exactly this creative struggle when we launched the project! Of course we knew Battlefront and enjoyed the game, but we always knew we couldn’t do these massive scale battles with lots of vehicles in Source with its limited grid size. BSP engines are just not the best choice if you want to do that. Still, there was the question of pacing. Since we were coming from the Jedi Knight series, we were used to a pacing that was far beyond that of HL2DM. All players had force powers, strafe-jumping through the maps at an insane force-speed, placing mid-air kills with the Golan Arms Flechette or pushing other players into bottomless pits with acrobatics or force powers. Looking back, I haven’t seen a single online shooter for the past seven years that required as much personal skill and coordination as Jedi Outcast.

On the other hand, Moviebattles was the exact opposite. It relied more on players having different abilities which you needed to combine in strategic moves to succeed. The mod had a substantially slower pacing. In the beginning, we were ambitious enough to try to please both target groups, as we knew they both existed. We planned CTF and TDM to be fast-paced, while using different mechanics for the objective-driven modes.

We soon had to find out though that we couldn’t get the strafe-jumping mechanics to work in Source the way that they worked in the id tech 3 engine, so we ditched the idea and went for the slow-paced tactical approach. Currently, we’re focusing our core gameplay around a set of base classes, which fill in a different role each. In our first public release, those will be four player classes, which we plan to extend with future patches. We’re gonna roll out more information on each class over the coming weeks.

Last Man Standing is an objective-driven mode, which plays differently on each map. Essentially, we’re establishing different missions on each of the maps with one team trying to achieve a certain goal while the other team is in the defending role. The objective can be getting an object to a target zone, reaching an escape point, destroying a generator or whatever the level designer can think of. The levels for this mode are rather small, so each round lasts for about three to five minutes. Once a player is killed, he or she cannot respawn. Thus, if an entire team is wiped out, the opposite team wins.

Thanks, I’m sure the artists will love to hear that! We always wanted to take as much advantage of the visuals as we could. Star Wars really lives through its amazing art direction and it’s a pleasure to rebuild famous key spots. To me, the visual component of the movies is all about the “used future” look with some rather shiny places like Kamino as a contrast, but all of the settings with a very romantic color pallet and lighting.

Star Wars has recently undergone the visual direction of stylized graphics. The Clone Wars or the new MMO “The Old Republic” feature comic proportions or cartoonish textures. We’re trying to capture the classic movie look instead. Technically, it’s mainly work with high-resolution textures and highpoly models used to bake our normal maps. It takes more work than creating a normal map from texture, but the results are superior. We also use color-correction on most levels to fine-tune the atmosphere.

In terms of framerate, the Source Engine works out quite well on indoor levels. It’s very scalable, so players with older computers won’t have a problem playing the mod.

We only ran into FPS problems on one map during playtests so far, which we’re currently optimizing. Since we’re still implementing some core features, we haven’t run playtests with more than twelve people though, so 24 player matches are something we’ll still have to explore.

It’s always a bit of both. We’ve had updates that broke all of the developer tools for a few months and we’ve had updates that opened new opportunities. For instance, we ran into severe jitter issues with the animation system due to some glitches in the lag prediction system in one of the older builds. This problem drove our coders and animators mad, until we found out that the orange box build came with an optimized animation system that swept all of our problems away. We sometimes ended up building various systems from scratch again. This did throw us back in time, but the results offered a better stability and performance.

The major problem until today is the awkwardness of the SDK tools, though. If you run an operating system that’s newer than Windows XP, you run into all sorts of problems and have to create workarounds just to be able to compile textures or models. Developer suites like UDK, Unity or the upcoming CryEngine 3 for indie studios are a lot more accessible, since you don’t have to use command-line tools. Gabe Newell recently said that Valve is focusing on the creation of better developer tools, so we hope the modding community will get a benefit from that as well.

We’ve made several bad experiences that have thrown us back substantially. Initially, we wanted to release live at FedCon in 2008, when we ran into severe engine problems with our projectile system. Then came a phase where we didn’t have any coders to progress with, until we were finally able to reboot the project in September last year. A lot of that time has been used for laying down new strategies and rebuilding certain components.

We’ve reduced the scope of the project for this very first release to create a target that we can actually achieve. As a result, the first release will only feature the Galactic Civil War era and four player classes. Right now, we’re in the process of migrating the last few core gameplay elements, after which the GUI will be getting a larger overhaul. At that point, we’ll be extending our internal tests to a larger user base and would also be happy to team up with communities who’d be interested in testing on their own Windows or Linux dedicated servers. Other than that, we’re busy giving the assets locked for the first release the finishing touches.

I wouldn’t want to name a release date at this point, though. Once we’re through with the coding-heavy stuff for our core gameplay, the Linux dedicated server and UI, we can think about a concrete date.

Thanks a lot, we hope we can proceed without further outages and hope to get some fresh people on board, as well.