A Decade of Norrath: 10 final things I wish I had known before building an MMO
Boy do we have a treat for you on this final installment of "10 things I wish I had known before building an MMO", in celebration of EverQuest's 10th anniversary. You may remember today's contributor from a recent artist spotlight. Kevin Burns is back to give us his retrospectively awesome list and share some priceless knowledge (and laughs) with any aspiring MMO developers out there.
As the days tick down towards the 16th, we're making preparations for that final day's contest. More details will be given on Friday, but you can count on it being worth the wait. Make sure to continue after the cut for Kevin Burns' extensive top 10 list and check out our complete countdown coverage thus far.%Gallery-46192%
Kevin Burns
Environmental/ Object Artist
Get out the machete, this is uncharted territory. As an environment artist, most of the unknowns (I wish I would've known?) of creating zones came with the new huge scale of Massively Multiplayer Online games. From the beginning, we knew EverQuest would be big, but we didn't know how big. And EverQuest definitely was, and still is, massive.
Here's a list of 10 things (in no particular order) that I wish I would have known 10 years ago, (actually 13 years ago would be better, at the start of EQ development).
MMO Level Design Layout – As we first started laying out dungeon zones, it was from the perspective of a single player or group. Very linear and focused on getting from the entrance to the boss and clearing out the dungeon. But we had to rethink a zone's layout as we had to adjust for multiple players/parties adventuring through and having the NPCs re-spawning more often. We would have to change zone layouts to include more alternate paths and offshoots, from the main route, to help spread players out more.
MMO Zone Size – Along with how a zone was laid out, the size of the rooms and passages had to be much larger as well. Nearly every space in the game that would be perfectly fine in a single player game had to be scaled up, from buildings for banks and merchants, to interior dungeon halls and caverns, and even outdoor spaces. Merchants or trainers had to be multiplied and spread out in larger rooms to accommodate more players trying to use them at the same time. Boss and raid areas really needed to be huge to hold multiple groups of players as well.
MMOs and zoning. In a traditional game, a player would load a new level, explore it from beginning to end, beat the boss, and load the next level (or get killed and restart the level). EverQuest was very different in how it handled zones. Many zones, especially outdoor zones, connected to other zones, creating a massive continent, which needed to be carefully planned out. Dungeons were even zones inside of other zones. This massive web of game-play areas required new zone layout ideas, and precise methods for creating a level's zone connection points. We soon found that we would have to get creative with zoning choke points, to block visibility of other zones, as well as having visual cues for leaving a level.
MMOs and exploring zones – In MMOs, you're only a quick in-game shout or online search to find anything in a zone. We had to learn to plan out zones where most players would already know how/when/where every was. In dungeon zones, the idea was to lay them out in an interesting way to promote exploration and the thrill of discovering the treasures and monsters throughout the level. A new player or group would have to explore a few different dead-end corridors or pathways, as well as be wary of monsters lurking around hidden corner, to successfully navigate through the entire dungeon. But in building a dungeon in an MMO, you have to assume that at least one member of any given group will know everything about the zone already, either from the online community, a guidebook, or other spoiler info, or a higher level character leading the group. So we would still layout the zones to encourage exploration, but we had learned to do it keeping in mind that someone in the group would already know every nook and cranny of the dungeon. Simple ways to keep players visiting different areas of zone were to spawn NPCs with unique quests or loot items.
The widespread variety of habitats in the world of Norrath – EverQuest has every type of real and fantasy landscape imaginable. Where most games have a central location and theme, and few have players globetrotting around to different climates, EQ has a whole world of environments from deserts, jungles, forests, mountains, tundra, and the underground, to all the places in between, and even into alternate dimensions and times, and even the planes of the gods. It was a very interesting challenge to create new textures, objects, and architecture of such a wide range when creating zones in EQ and keep it feeling like the same game universe. A world builder would go from creating a forest zone one month to a desert zone the next, and have to essentially start from scratch. As we created more and more zones, we continued to streamline the process and work more cohesively on between similar environments shared by different artists.
New Tools, Level Creation – While most of the art assets were created with Photoshop (textures) and 3DSMax (outdoor zones, objects, characters, weapons, etc.), the first interior dungeon zones where created with an in-house (2D) level creation tool. 2D because zones were defined from the top down view, similar to drawing them out on graph paper, then each room or area was assigned a floor and ceiling height, and then exported in 3D. The tool and exporter (affectionately known as Ike and Tina, respectively) were created from scratch, and we had to learn them and use them as they were being built. This had lots of interesting little quirks like "...save often, there's no UNDO in the tool yet". Eventually, the 2D dungeon level creation tool was retired and interior zones would all be made in 3DSMax.
New Tools, In Game – More custom tools. Getting all the art assets in to the game also required creating and learning more new tools on the fly. A custom object placement tool was needed to add all the static objects (tables, chairs, trees, rocks, and other décor) to the zones. Also, an in-game avatar tool was soon needed to add the more interactive elements to a zone, such as lighting, sounds, particles, and NPCs.
Crunch Time. Crunch time at the end of the development cycle was certainly not unknown. It was pretty common at the time for most games to have at least of a couple of weeks of crunch time (and probably a solid month or two was more the norm), for polishing and bug fixing. But with the massive scale of EQ came massive crunch time. It seemed like forever. Fueled by pizza and takeout food, Mt Dew and coffee, and the desire to make a new, fun, and entertaining virtual world, we spent the better part of the 3rd year developing EQ locked in crunch mode. As much time as we put into the game before it originally launched, there was still a lot of content we weren't able to get in. Eventually, it would find it's way into the game through numerous expansions and live content.
Learn by doing, and try again. A lot of the art assets and some of the first levels were actually re-created a few times over, as we figured out what worked and what didn't work in a 3D MMO. Zones that were built in the last year of the dev cycle were more refined than those built in the first year. Some of the original zones had to be rebuilt again and again. Some areas of certain zones had to be rebuilt, in ways to block visibility of areas where the poly count would get too high. Other areas needed "invisible wall" objects added to keep players out of areas where they might get stuck. Every new level the environment team created led to learning more about what worked and what didn't, and refining the world building process with better ideas and updated toolsets.
System Specs, a moving target. I think everyone on the team, artists, designers, programmers, managers, and marketing, would have loved to have known the exact system specs of the game from beginning. As with most computer games, predicting the system specs of your target PC, 2-3 years before release, is a difficult task. EQ was too large to go cutting edge graphics like the FPS's of the time, but still had some pretty hefty PC requirements because of all the content we wanted to deliver. Yes, many people had to buy a new 3D video card, if not a new PC, to play EQ when it launched. As we continued through the dev cycle we were able to narrow the scope of our target machine, and adjust our art requirements (texture size, poly count, etc) better towards the end.
So, in the end, as we charted new territory, building 3D environments in the emerging MMO genre, we certainly learned a lot. Some practices are still alive in creating MMO game zones and objects today, while most have improved by leaps and bounds.
