// overview
Six promising young creators from developer Tri-Ace sat down for a candid interview during production. Topics ranged from unknown design details and behind-the-scenes development stories to the team's hopes for what the game could become.
"I think we were expecting something lighter but that's Tri-Ace for you."
Designer · Art Director
T. Takayashiki
高屋敷 哲
Overall art direction. Oversees company graphics and manages art-related staff.
Planner · System / PA
T. Asanuma
浅沼拓志
Handles main scenario, and PA & setting planning. Worked as subordinate to the main scenario lead.
Graphicer
H. Ichihashi
市橋秀明
Monster design lead, focusing on motion and animation.
Programmer
T. Ookushi
大串輝騎
System programmer. Responsible for field-map systems and battle-related development.
Programmer
N. Kamei
亀井尚久
Camp screen programmer. Covers Edifice, skill displays, and character relationship development.
Programmer
M. Nishimura
西村友樹
Item Creation and communication-related development. Also worked on items and some battle content.
This was originally going to be a Game Boy title, wasn't it?
Takayashiki: Originally, yes - we were making it for Game Boy, aiming for something more compact. The concept was basically a step up from Star Ocean (the original, "SO1"), taking SO1's adventure story and adding features like the PA system from SO2. But we ended up turning it into something bigger.
How did it grow from there?
Takayashiki: SO2's PA system had been fun to make, and we wanted to take it further - really let players experience Private Actions. From a direction standpoint, we also wanted to try and show what a full, proper Action RPG could look like on Game Boy. Then, as development went on, the scope kept expanding... and here we are.
How many people are on the development team?
Takayashiki: About twelve people total. At peak we had around sixteen mid-development, but the base team is twelve. Honestly, for a Tri-Ace title, that's small. (laughs)
What was the concept you were aiming for?
Takayashiki: At first we tried to keep it simple - not "let's go all out." But once real development started, we kept adding things and things... it ended up being very Tri-Ace. (laughs) I had planned to release it in a year, but here we are.
Asanuma: The scenario was supposed to be lighter too. But Main Scenario kept adding depth and it snowballed.
SECRET FILE 01
~ Edifice ~
01
Edifice - Original Design
Edifice is fundamentally a water-based mechanical lifeform. Its body structure is originally humanoid and its purpose was to serve as a leader or vanguard-type model. Edifice has the ability to transform its body into something soft and fluid - resembling a sea creature - in order to travel underwater. In-game, however, this transformation is never shown. One of the more striking early design concepts was that the Edifice body was to have been translucent, revealing muscular structures visible beneath the surface.
Early Settings & Cut Content
Were there settings that changed drastically from the original plan?
Takayashiki: The SO2 setting has three PAs (in that game) and it was originally going to use those as a base (for the story), and expand outward. Outside characters were also planned as PAs - for example, Valkyrie Profile's Lenneth was at one point going to appear as a PA character. We ultimately decided that it should be a proper character.
Asanuma: The original cast was planned as three characters - Rena, Claude, and Ashton. Then it got revised to just Rena and Claude. But after everything expanded, the final lineup includes Precis as well as Celine. (laughs)
Were there characters who were completely cut?
Asanuma: Raspos was one. She was originally going to be a female character. One of the big early goals was to have at least one girl per character type, so Raspos and Ashton would be a matched pair - but ultimately she was cut.
How did you handle the transition from Game Boy to GBC?
Takayashiki: Full GBC support, yes - that was the plan from the beginning. It was a big challenge. Given Game Boy's constraints, we had to express character movement and action in a way that read clearly. We went back and forth on how big to make the character sprites. For a normal Game Boy game, 2 sprites per character is standard, but we ended up using 3.
And the battle system on that hardware?
Takayashiki: Monsters were extremely demanding. We had to work hard just to make them move well - a single large monster nearly didn't fit. In the end, we used Pandemonium-style giant enemies. The monster that caused the most problems was the one for which the hitbox alone would barely squeeze onto screen.
Tell us about Edifice - its true nature, where it came from, what it is.
Takayashiki: That line from SO1 - Edifice - was something we had wanted to explore for a long time. In SO2, a Moon - a massive body - was moved towards a planet on a crash trajectory. The Edifice is essentially that Moon's AI. It drifted through space and ended up landing on the new planet after the events of SO2.
So Edifice is not originally from the Edifice planet?
Takayashiki: Correct. It came from space, crashed there long ago. The civilization it created - Atlantis-like in nature - was basically born around its presence. We wanted to tie the SO1/SO2 mythology into SO:BS properly.
Asanuma: In terms of the game's storyline, Edifice is the "mother computer" that controls the civilization of the planet. It is a biological machine - a living computer. Its true form is something like a giant sea creature with psychic abilities and communication power. The strange events affecting the planet can be traced back to its awakening and the large-scale environmental changes it's causing.
The "Rival" character - what role did they serve?
Asanuma: She's basically 1–3 levels ahead of the party at all times. She serves as someone who is constantly ahead - but not by enough to make them feel unreachable. She's not evil, exactly. She's an Edifice-type character who is trying to do something the party doesn't fully understand yet. In a way, she's the most honest character in the story.
SECRET FILE 02
~ Muninku No. 3 ~
02
Muninku No. 2 - Cut Character
A self-propelling space patrol drone originally developed by Edifice. Muninku No. 2 was designed to activate when the planet came under external threat, launching from the planet's surface and engaging the threat directly. Its original purpose was to scan, assess, and destroy debris and objects that threatened the planet from orbit. It was cut before final release. Muninku No. 3, which does appear in-game (As the Atlas Enemy), uses Muninku No. 2's right shoulder as its main gun barrel. The body underwent multiple design revisions, making it somewhat different from the initial early-setting version.
// difficulty & characters
The difficulty feels high. Was that intentional?
Takayashiki: That's always how it goes - during development it feels fine, and then actual playtests reveal things. The target demographic shifted a bit too: it was initially aimed at middle-schoolers or high-schoolers, but became more oriented toward experienced players. The dungeon hints turned out to be a big point of friction.
Kamei: Silence status being too easy to apply was a problem. Having 100 control inputs needed with only a few buttons was another. (laughs)
What was the internal motto around difficulty?
Takayashiki: "If 1000 people play and even one person clears it, that's fine - and someone inside the company will definitely manage it." (laughs) We probably pushed it too far.
How were the playable characters decided?
Takayashiki: The main lineup: Rena and Claude, was fixed from the start as a given. Ashton was also a priority. After that, it came down to what type coverage we needed mechanically. The goal was to have at least one close-ranged fighter, one mage, one support, and one flexible type. Precis fulfilled the flexible role in a way that no one else could.
Any characters you're particularly proud of?
Asanuma: Honestly, Ashton. His character arc with Creepy (Gyoro) and Weepy (Ururun) is something I'm fond of. Using him properly - leveling his Special Arts, understanding the timing - is genuinely satisfying when it clicks. He's the one who rewards player investment the most.
Were there major balance problems?
Takayashiki: We tried to keep balance even from early on, but it was hard. At one point, the difficulty relationship between dungeon and town was extremely off. The final version tries to ensure even young players can enjoy it with proper use of Items and Crafting. I personally believe you can beat the game solo with a single character if you play it right - though I wouldn't recommend it. (laughs)
Asanuma: Debugging was a nightmare. When we finally did a full playthrough with all skills maxed out, one person had soloed the entire game with Ashton. That was... a moment. (laughs)
Was Item Creation hard to balance?
Nishimura: Extremely. IC was the last major system to be finalized - it kept getting adjusted until very late. The goal was for it to feel rewarding without trivializing combat, but a maxed Crafting skill plus Duplication makes things very different. We tried to ensure the game never became entirely dependent on it - it should augment, not replace.
Dragon Warrior? Blue Sphere?
Kamei: Dragon Warrior wasn't something we took directly - it's more that the action-RPG genre broadly shares some vocabulary. (laughs) Blue Sphere was an item whose properties were tied to a water stone on the planet. Finding it legitimately required knowing where to look. It was not supposed to be easy.
Takayashiki: We wanted items with interesting lore behind them. Not everything should be findable in one run. Some items exist purely to reward replay.
SECRET FILE 03
~ Edifice (Mother Computer) ~
03
Edifice - The Mother Computer
The ancient civilization responsible for shaping the planet of Edifice was controlled by a central mother computer - Edifice. Its true form, an ancient artificial structure of enormous scale, resembles a great worm or larva in battle - though this is only its combat configuration. In the storyline, seeing its full form is possible, and Grand Knott has a special code that allows the player to access its complete design data. This is the largest and most complex enemy model in the game, requiring special rendering techniques to display correctly on the GBC hardware.
// broader audience & closing
Reaching a Wider Audience
Is there anything you wish you could have done differently for newer players?
Takayashiki: Honestly, if I could redo it I'd adjust some of the field hints. There are moments where the game doesn't communicate clearly enough what you need to do next, and players just wander. That's not fun difficulty - it's unfair difficulty. Those are the moments I regret.
Asanuma: The saving and equipment interface could be more accessible, too. Bloody Armor is a great example - it's excellent in the right hands, but almost nobody figures it out without looking it up. If even one thing from Item Creation could be unlocked automatically early on, it would have opened things up.
Next steps for the series?
Takayashiki: Not an RPG. (laughs) Ideally something shorter - a project where the whole team isn't consumed for years. But of course, this is Tri-Ace, so whatever we make will probably end up bigger than planned.
A message to players?
Ichihashi: Please 100-point the game. It took a lot of effort to make it completable, and we'd love to know someone out there did it fully.
Ookushi: Above all, enjoy the characters. There's a lot to them - push their bonds as far as they go.
Kamei: If you've played it once, play it with a partner. This game has a communication feature precisely because it's better shared. The things you discover together are different from solo play.
Asanuma: Players who finish may find themselves wanting more. If you do - we think that's a very good sign. There's a lot hidden in there. Keep exploring.
Nishimura: Don't neglect Item Creation. It's not just a side system - it's what separates a first run from a real playthrough.
Takayashiki: We're still all working hard. This one came from the heart - we hope you can feel that.
Communication play feature?
Nishimura: The communication system was built with Item Creation as its backbone - trading via the link cable was designed to let you share crafted results. Very early on it was planned as a fully fleshed feature, but time constraints pushed it to what it became. We still think it's worth using if you have two copies.
// source note
Translated from Star Ocean: Blue Sphere Final Guide Book scan, pages 268–272. Interview conducted on August 16, 2001, 15:30–17:30 at Enix's offices. Some lines condensed or paraphrased for readability. Speaker attributions restored from context where scan quality degraded. All content copyright Square Enix / Tri-Ace.