XiaNaphryz
LATIN, MATRIPEDICABUS, DO YOU SPEAK IT
Just got my copy in the mail, this month's postmortem is on FFXIII and written by Motomu Toriyama and Akihiko Maeda. Not going to post everything here from the 4 page article, just the bullets and a brief snippet from each of them from the usual postmortem sections.
It's a decent read, but reading through the lines I suspect what they detail about the game's development is probably on the understated side, given what they ended up listing for the "What went right" section.
As someone who's both been on and observed several development teams in Western gaming companies, it's always fascinating to hear about how Japanese developers operate. The Location Manager point is a good example - this would just be covered by someone on the production staff over here, usually an associate producer along with an assistant producer or two.
If you want the whole thing either snag a copy (print or digital) or just wait for the article to be eventually posted on Gamasutra.
WHAT WENT WRONG
1) LACK OF A SHARED VISION - Final Fantasy XIII was first introduced through a concept trailer shown alongside the announcement of the Fabula Nova Crystallis project at E3 2006 (Fabula Nova Crystallis represents a suite of games and other entertainment media related to Final Fantasy XIII). The trailer was merely a visual concept and we had not yet created anything playable at that point.
<snip>
2) THE UNIVERSAL ENGINE AND NARROWING DOWN THE SPECS - Another issue was the universal engine. Because we were so focused on creating an engine for next-gen hardware that could be utilized across all platforms, we made the mistake of trying to accommodate every single project that was in progress at the time. In hindsight, it should have been obvious that it would be impossible to fully satisfy all of those needs.
<snip>
3) GETTING STARTED ANYWAY - As all of this was going on, the staff involved directly with the actual data construction had no choice but to start working before the specs were finalized. Their main concern was that they could not be able to keep up with the schedule if they continued to wait for final decisions.
<snip>
4) THE LIMITS OF THE TRADITIONAL TEAM STRUCTURE - As the project's scope increased, the traditional development style of dividing the team into specified roles, such as character modelers or texture artists, started to present issues as well. This problem of over-specialization presented itself in each discipline. The biggest problem was that the project became bloated with the increase in staff within each department. And because roles were so specific, the communication flow became faulty and information was not being shared properly.
<snip>
5) INTERNATIONAL PLAYER TESTS THAT CAME TOO LATE - Even before the current generation of consoles was introduced, it was obvious that the game market of the West was gaining momentum, and we couldn't ignore it. The sentiment that stood out the most to us at the time was the increasingly harsh criticism towards JRPGs, Linearity and command-based battles were tow of the features being perceived negatively. This was something that the team was very conscious about, and there were concerns about whether JRPGs would still be accepted in the West. Because Final Fantasy XIII's mission was to succeed worldwide, we could not ignore this issue, as we felt it could deeply affect the future of the franchise.
Around the same time, we were experimenting with Western development methods, and there were talks within the team of global focus groups, which we had rarely conducted with previous projects. At the same time, Square Enix set up international focus groups for certain titles, including Final Fantasy XIII. Unfortunately, we were already quite far along in development, and knew it would be too late to implement most of the feedback from the player test sessions.
<snip>
WHAT WENT RIGHT
1) REALIZING A SHARED VISION THROUGH THE DEMO - Even at a late stage of development, we did not agree on key elements of the game, which stemmed from the lack of a cohesive vision, the lack of finalized specs, and the remaining problems with communication between departments,
What enabled us to conquer this line of seemingly endless conflicts fwas the development process for the Final Fantasy XIII demo, which was included in the Japan-only Blu-ray version of the animate film Final Fantasy VII: Advent Children Complete. The demo was not in our original plan, so we had to make adjustments to overall schedule to accommodate it. Whatever effects creating the demo had on the schedule, once it was complete we realize it was just he panacea we needed.
<snip>
2) CLARIFICATION OF ELEMENTS AND PROCESSES THROUGH DEVELOPING THE DEMO - The demo brought together all data, development of which was previously uncoordinated, clarifying many elements and significant speeding up the process of determining the remaining specs.
<snip>
3) CREATING THE LOCATION MANAGER ROLE - Although the schedule was now working, we began to realize that we were unable to keep up with the sharing of information within the traditional team structure. On order to resolve the issue, we created a new role that did not exist in our traditional development environment the location manager, who would function to bridge the gap between different departments.
<snip>
4) RESOLVING THE UNIVERSAL ENGINE DILEMMA - The issue with the universal engine was seen as something that would affect the progress of all related projects within the company, and in the end, it was decided that the needs of the flagship title, Final Fantasy XIII should come first and foremost.
<snip>
5) NARROWING DOWN POLISHING POINTS THROUGH FOCUS GROUPS - Through the focus group we conducted (mentioned in the "wrong" section), we found that, contrary to expectations, the game was received very well by Western players. Also, both Japanese and Western players place emphasis on the story and battles, meaning that the style we focused on with Final Fantasy XIII was accepted after all.
<snip>
The conclusion to the postmortem is very simple. I learned that we must first create something tangible and playable to share a game's concept and confirm the actual specifications to create an environment in which precise decisions function within the process. Only then can you proceed with the development of such immense amounts of data. Furthermore, up until this point, we had taken a very traditionally Japanese "Square Enix method" based on individual craftsmanship. Now we are trying to incorporate as much as possible form the development style of studios overseas into our system.