[In this postmortem taken from the December 2009 issue of Game Developer magazine, iconoclastic developer Double Fine Productions outlines the complex and at times daunting process of developing its action/strategy rock god epic Brütal Legend, taking in everything from legal troubles with Activision to tool development and everything in between.]

If the adage "go big or go home" applies to any software development effort, it applies to the making of Brütal Legend. As we did previously with Psychonauts, Double Fine once again bet it all on innovating -- this time on a game borne from the Full Throttle side of Tim Schafer's mind. Brütal Legend is a molten, balls-forward, third person, open world, strategic action-adventure interactive ride into the very soul of heavy metal.

The development story started out simply enough. After shipping Psychonauts, Double Fine created a collection of concept work, a pitch document and a game trailer intended to capture the spirit of Brütal Legend. Most publishers we spoke to were interested in the game concept, but their questions commonly indicated that they didn't understand where we were headed with it.

Questions were posed to us, such as "Why heavy metal? How about rock, or country, or hip hop instead? Why would you want to play as a roadie? How about playing as a rock god?"

One publisher, Vivendi Universal Games, did not ask these questions in the pitch meetings -- or in any meetings. They understood the game for what it was, and signed it for what they knew it could become.

We started development by focusing first on the multiplayer mode of the game -- our thinking was that since we'd never made a multiplayer game before, figuring that out would be our top priority. It took 16 months to do so.

In Rocktober 2006, we delivered a fully playable Ironheade vs. Tainted Coil skirmish to our publisher. At Vivendi's request, we then focused on the single player campaign, expanding its scope well beyond the initial design.

This is the period in which we added the voices of Jack Black and a host of other celebrity talent to the game, as well as other enhancements that solidified the vision for the campaign experience that we ultimately shipped. In June 2007, we delivered the first meaningful portion of the single player game and also promptly admitted that all of the changes we had made to the game content put us way behind schedule.

Our first schedule revision extended the project by ten months, the second by another seven. Initially scheduled to be released in May 2008 under the Vivendi Universal Games/Sierra banner, Brütal Legend finally shipped on October 13, 2009, published by Electronic Arts.

Double Fine encourages innovation, but that drive also means we can't always rely on previous experience to predict how a feature or an approach will turn out. On Brütal Legend, the practice of continuous iteration and concept refinement led to a number of prototyped ideas, many of which survived to ship in the final game, but just as many of which were left to digitally fossilize in the annals of Perforce.

Whittling down to ten the list of things that went right and wrong during the development of Brütal Legend presented a considerable challenge. Here are some lessons that were the most surprising or impactful.

What Went Right

1. Pushing Creative Limits

Brütal Legend was to be the interactive amalgamation of the over-the-top ridiculous (yet deadly serious) world of heavy metal. We were reverent fans of the genre and felt it would be an honor to bring that world to life. Brütal Legend began with a simple list -- a game that embodied everything that could be found on a heavy metal album cover: chrome rivers, pools of blood, volcanoes, caves, fire-breathing metal beasts, laser panthers, bladehenge and beerhenge, dominatrices, latex and chains, disembodied undead heads...

To that we added the core gameplay. We wanted to make a brawler adventure game, where the player was a heavy metal roadie who evolves into a rock god over the course of the game. Brütal Legend on the Xbox 360 and the PlayStation 3 would have the brawn of an action game and the elegance of an RTS.

We learned early on in our relationship with Vivendi that RTS was a naughty word in the console space, so we stopped calling it by that name and, by extension, so did Electronic Arts -- positioning the game largely as an action title in the marketplace.

We wanted our RTS to exploit the consoles' advantage; putting the player in the center of the action. We wanted to give the player intuitive control of a character that could perform a variety of badass movements and abilities and also allow the player to personally bond with that character. And we wanted that character to command dozens of masterfully-dialogued troops simultaneously.

One of our biggest challenges in solving the RTS accessibility issue was squad orders. It took numerous attempts and countless focus tests, but we ultimately decided on a simple unification of the orders interface, wherein the AI behaved as the player would want them to on the battlefield.

The player would have four orders:

1) "Follow," where Eddie gives his troops commands to move to a specific location and possibly attack, allowing the order to be given only within "shouting range," a relatively large distance around Eddie that did not encompass the entire map -- solving the forced (ignore path enemies/obstacles) and non-forced (engage enemies along the way) issue by making the "follow" order non-forced when Eddie was near his troops and forced when Eddie was far away.

2) "Defend," where Eddie could command his army to stop moving and hold position, aligning them in the most advantageous manner (melee up front, ranged behind, support in the rear) facing the camera.

3) "Move," where the migration would be forced until the army is close to its destination, at which point it would engage nearby foes.

4) "Charge," a non-forced move to the enemy that is closest to the average position of all nearby enemies if your army was not attacking, and a forced move to the attack position even if that meant disengaging from their current activity if your army was attacking.

The Double Fine incarnation of a console RTS occurred to us not in an early pre-production meeting, but over years of painstaking iteration and reinvention and rework. We tested our progress in periodic "Mandatory Hour of Fun" sessions, where the entire team played the latest build and then met as a group to discuss what was working well and what was frustrating or could be made better. This open forum for the exchange of ideas as well as the momentum for continuous iteration fueled profound changes to the core game mechanic over the course of development.