We recently had the opportunity to sit down and interview independent developer Derek Smart, whose veteran works date back 14 years, to the somewhat controversial Battlecruiser 3000AD. We spoke to him about his upcoming MMO Line of Defense. We’re taking a bit of a different approach here, as this interview is far more technical than ones we’ve performed in the past. Let us know what you think, and if you would like to continue to see more content down a technical avenue.
AO: I’ve noticed that you guys are using the Havok Vision engine. What sparked that decision? Has your previous work utilized the same engine?
Smart: We have always developed our own game technology in-house; only licensing middleware such as audio, networking and other miscellaneous tools. For Line Of Defense we started doing the same thing when we started the project.
The original plan was to build on the in-house custom game engine that powered my last two games (All Aspect Warfare & Angle Of Attack) released in 2009.
After careful evaluation of that plan, it was determined that it would take too much effort to bring that engine up to speed and have it targeted for a game due out in the 2012 time frame.
So, we started developing a brand new game engine from scratch. After several months of doing that for the game’s prototype, I determined that it would take too much effort to complete, remain competitive – and release a game within a reasonable timeframe. Things weren’t progressing fast enough and so I scrapped over six months of work and we started from scratch on the engine development.
Prior to that, I had looked at various engines – all of which were either too expensive for what they could do, or they fell short of what I wanted to do in the game. For example, some engine developers tend to tailor their engines around a specific type game (e.g. FPS). So using such an engine for a game that was to feature various types of gameplay (e.g. vehicles, aircrafts) like my own engines did, was a non-starter.
But one engine stood out. The engine was Trinigy’s Vision Engine which later became the Havok Vision Engine late last year. This was the first time that we’d ever used the engine. The idea was to pick one engine, take it to the limit (while driving the engine developers mad with requests, suggestions, bug fixes etc) and then adopt it for all our future games (e.g. we have Galactic Command Online following LOD) going forward. This way, we didn’t have to build an engine, nor re-invent the wheel.
The cool thing about the Vision Engine was that it incorporated a bunch of middleware tools which could either be used (side loaded) or completely disabled. These include audio, networking, GUI, scene lighting, physics etc. Some (e.g. audio, physics) were free, while others (e.g. the insanely expensive Scaleform GUI) needed additional licensing. In the engine, we licensed the middleware ones that we needed.
In the end, we completely scrapped the prototype game engine we were building and went with Vision Engine and other middleware which included Havok Physics (which replaced Physx), Silverlining atmospheric lighting, Triton water rendering, Iggy GUI (because I’m not mad enough to pay $50K+ for the only GUI middleware that Vision Engine natively supports) and a slew of others.
Porting all our pre-existing content over was not as horrendous as we thought; though we did have our moments with our terrain scene which needed us to write a custom exporter in order to bring it into Vision Engine.
In the end, this was a good decision and I expect that my days of building game engines is over because all our custom additions to Vision Engine have pretty much given us a custom engine which supports all the features (and then some) for which I used to build my own engines in-house.
So now we’ve got first person infantry, aircrafts (including fast moving jets), vehicles, space(!) and other nifty things we incorporated – including a seamless indoor renderer.
So I expect that we’ll be using this custom engine for all our games going forward. And since the license fee is within our budget,coupled with the fact that the Havok name means that it will be around for a long time, I think this was a good decision. Now after using it to build our own custom game engine, we’re focused on actually building a “game.”
Your server system is impressive! The ability to sale between 2048 – 23k players on a single server session is a large amount, but I believe it lends for a greater experience for the players by not having to use instances. Why did you choose to not use instances?
Smart: It’s not as impressive as it sounds to be honest. A few years ago, yeah, maybe. The idea here is that I didn’t want a session
based (see Tribes Ascend) game because the game was built for massive battles in various simultaneously updated environments.
By the same token, this is not a game whereby we wanted “millions” of players which in turn would mean lots of servers. There is no glory in touting those numbers if gamers are either not having fun or they’re suffering from performance issues.
Plus, how many gamers actually do end up encountering each other in large games? Life happens – and means that it is difficult to get all your friends to go see a movie with you, let alone go play a video game. So doing it this way, we keep it small and somewhat intimate so that you can jump in, play for a bit and leave. If your friends are around, so much the better. You need millions or even thousands of players in order to have a fun and engaging combat experience.
With that, I asked my old friend Martin Piper, the author of the exceptional ReplicaNet networking middleware used in my past games, to come up with a custom architecture which would a) not end up with us doing a session based game b) give us at least 256 players per scene but with the ability to scale up or down as needed.
So if you did the math for a single server session, you end up with 256 (clients) x 13 (the total numbers of space, planetary, indoor scenes) gamers in a single server “instance.” This is manageable because we can scale up or down without having to roll out a ton of servers, then have to deal with server issues once the population leveled out over time.
From the onset, I neither wanted a session based game nor an instanced one. They are similar in architecture and neither is suitable for this game. So we went with this custom design instead and which allowed us a lot of leeway. Especially since we can also deploy third-party cloud servers without ever needing to roll out expensive hardware which would be obsolete at some point in time.
When browsing through the assets catalog I’ve noticed that you guys have quite a variety of vehicles and buildings. How large of a team was required to build such an asset database? Additionally, how long have you guys been working on the title?
Smart: The title was an idea of mine since early 2010 due to how All Aspect Warfare (released in late 2009) was received. Though Galactic Command Online – a massive game in the same scope as my previous niche Battlecruiser and Universal Combat games – was
my next follow-up title, I decided to do a parallel title given how long GCO would take, the ever changing industry climate etc.
The content team was quite large and we have guys from various countries. In fact, we have so much content that it was only as recently as this past April that we were finally content complete!Development on LOD didn’t go into full swing until mid 2010.
And just recently, I decided to go back and improve on some of the weapons because they were done almost two years ago under different engine specs and now they needed to be brought up to the standards of the new Vision Engine, current gen quality etc.
I see that you are currently signing beta testers. Is the beta open for all at the moment? Are the servers live? If so, what has been your greatest surprise upon doing the beta?
Smart: We don’t have the resources – nor the patience – to do an open Beta. So a closed beta test (CBT) is what I decided to go with. As complex as my previous games were, one thing I found is that the smaller and tighter the testing group, the more valuable the
feedback. We don’t have resources to waste on freeloaders who aren’t going to contribute to the development and improvement of the game.
As of this writing, the CBT servers are not yet live, though we already had the sign-ups for over a month. Once they go live, we’re going to make the selection at random (unless you’re in the media or are in our pre-existing game tester list from previous years) and flip the switch.
As the game is due out by the end of 2012, we expect that the CBT will go live very soon.
Your studio, 3000ad is located in South Florida. What is the development environment like down there? Do you believe working in Florida, as oppose to California, which seems to be the gaming Mecca, has affected your development in a positive or negative way? How so?
Smart: Well game development doesn’t really rely on location all that much. Especially in this day and age. I have contractors all over the US and the world; so the environment doesn’t play into it all. That’s the joy of being an indie studio running on the lean and mean.
Basically, you could work from your garage, at thirty thousand feet in the air, or on an island.
If you can’t wait to get your hands on Line of Defense you can sign up for the closed beta now, at the developers site.