Interview: Tim Trzepacz on Princess Maker 2 and the struggles of 1990s game localization

Tim Trzepacz has been involved in the video games industry since MicroProse’s Sega Genesis version of Pirates! Gold was released in 1993. He’s worked on games like Ratchet and Clank, James Bond 007: From Russia with Love, Magic Knight Rayearth, Lunar: Silver Star Story Complete, Silent Hill: Origins and Silhouette Mirage. We talked to him about Princess Maker 2, which his company SoftEgg attempted to release outside of Japan between 1995 and 2002.

Michibiku: Thanks for talking with us, Tim! Could you start off by telling our readers a little bit about yourself and your work with games localization?

Tim Trzepacz: I’m Tim Trzepacz. I also make music under the name Timon Marmex and maybe some other names. I have a degree in Electrical Engineering from Virginia Tech, and have been in the game industry for over 20 years. My first commercial game was Pirates! Gold for the Sega Genesis. I’ve worked on something like 20+ titles in my career, including the Ratchet and Clank series, Silent Hill, James Bond and other things you might have heard of.

Early in my career, when I was still working for MicroProse, they were having some hard times and disbanded the localization department. So, I quietly had all of their Japanese game magazines forwarded to myself. And that is how I learned about Princess Maker 2. Most of the games were very porn-y and only ran on PC-98 systems. But there was a DOS-V version of Princess Maker 2. So I suggested to some friends that we try to license it and localize it for the US. Since I was still working at MicroProse, we put it under my friend David Leary’s company Adventions. We did everything we set out to do except finding a distributor.

We think of games localization as sort of a separate entity today, but it seems you’ve managed to dabble in it while doing primary work elsewhere. How does that make the process different?

TT: Well, I did Princess Maker 2, and then worked solely in localization for 4.5 years at Working Designs. At Working Designs, we had a bit more division of labor, since we were doing it every day. On Princess Maker 2, just about everybody worked on everything. Leary didn’t do any code, though.

At one point I was trying to do a fan translation of Sakura Wars without source code, but I decided that I didn’t want to do my day job twice, so I stopped. It was going to be tricky without source code. I had managed to extract all of the text, but only as bitmaps. So I couldn’t use machine translation or dictionary lookup on it. If I had a license for it, I could have had the real text for a proper translation. Changing the text engine without rebuilding the whole game would also be very difficult.

Institutional support aside, which do you think works better: everyone working on everything or people dedicating themselves to specific roles?

TT: Both have their advantages and disadvantages. If you have really motivated and talented people who can sit in any seat and make magic happen, then that is great. But a lot of people aren’t good outside their specialty, or they just aren’t that motivated, and then it is best to have a set role for them.

Princess Maker 2 is a gigantic game that’s very text-heavy. Where did you start and how did you proceed through the translation process?

TT: I think we started by trying to get the license, which we did by asking Robert Woodhead of Animeigo if he had any tips or contacts. He ended up representing Gainax in our early talks. It was very slow. Once we had an agreement, they were able to send us code and data.

Was that code organized and easy to navigate? We know things were less standardized those days.

TT: The whole thing is written in a custom “BASIC-like” language called “PM-BASIC”, which is implemented in 8086 assembly language. The text was interspersed with the game logic in that. That is generally a no-no to have text and code together, because it means each language has to have different versions of the entire code. But in this case, since we were just doing English, it was fine. Crawling around in assembly commented in Japanese was interesting too! Although I had a lot more of that later with Working Designs.

Is that why you also had to do both programming and translation work when preparing that localization of Princess Maker 2?

TT: Well, you always have to do a bunch of reprogramming simply because the text engine requirements for Japanese are very different than for English. We had to do some very interesting changes just to get statistic names to fit in the boxes!

Was “Temperament” one of the stats that gave you trouble?

TT: We made our own font, and for statistics names, the word would have several characters that ran together to make the word. I don’t really remember what they all were, but you can see some of them which have two lines of text crammed in and such.

Were you also responsible for programming those Cheat Shops that were included in the game?

TT: I don’t remember adding anything like that. I think those may have been in the original source. Perhaps they were disabled and we re-enabled them for testing? I vaguely recall [Chris] Nebel might have found those. I don’t think we added anything, so it was probably part of debugging that we turned on. Remember that the version that got out was not intended to ever reach the public.

With niche games — the kind like Princess Maker 2 and the kind that Working Designs tended to pick up — is the process generally tougher, and how? We’d guess that these smaller teams are less prepared for the task, and not just in the sort of business ways that tripped up Princess Maker along the way.

TT: The problem with Princess Maker 2 was that the U.S. was not quite keen on anime-style art yet, and our game industry was fixated on Doom and FPS games at the time. Also, it was in the transition to Windows 95, and nobody wanted a DOS game anymore.
For a small team, it was easy to produce, but we weren’t set up to do distribution or promotion. Our agreement was to find a distributor, and there were none. Our agreement with Gainax had already run out when Intracorp contacted us. We had to negotiate a new agreement, which was a 3 way agreement. That was what prevented us from taking it over ourselves when Intracorp went under.

Were you at all concerned about the content in Princess Maker 2 after picking it up and getting the license? There are moments when it can go from something appropriate to all ages abruptly shift to something more mature. Also, it has religious overtones.

TT: There were a couple of scenes which we thought were inappropriate, and we changed those in the way that we felt was most respectful to the content, in our version. I don’t know what they did in this new version. Japanese culture is different than ours, and has different taboos. I didn’t think of it as a game for little kids, so I wasn’t worried about it too much. I don’t think the religious stuff was anything anybody should get offended at, so I didn’t worry about that at all. When we licensed it to Intracorp, they were pushing us to uncensor stuff we had changed. We were very against that. I think that was the folks at Ignite that wanted that.

Princess Maker 2 has quite a few pop-up conversations involving rivals, fortune-tellers and peddlers. Since the version of the game you translated didn’t have voice acting that you could turn to for possible personality cues, how did you translate these segments to ensure characters had unique voices?

TT: Bryan Buck did a lot of that rewriting, as he was also the main translator, which meant that he was aware of the nuance of the language in the original. We were all avid readers of science fiction and fantasy, so we were all familiar with the various tropes of the genre. And I had run the VTAS, the animation society at Virginia Tech, so I knew a bunch about anime and manga tropes and character archetypes. Takami Akai’s art was very good, very compelling, and gave us a good idea of the characters. So between the original text, the art work, and our knowledge of the genre, we were well prepared to write the characters in a way that would really bring that character out… in my humble opinion.

I felt like there was a lot more room for narrative in the Princess Maker games that I didn’t see in Princess Maker 2. I think I would have liked to have worked on a new Princess Maker game at some point, but it was not to be. When we made the real demo version of Princess Maker 2 — that was not released by Ignite — it had a two-year time limit. We wrote a little story where the god that gave you the daughter came back to collect her, because there was a war in heaven. I thought it made it interesting that other things were going on.

Did the demo have any extra endings, depending on how you did?

TT: Now that I think about it, I think the God came down to collect “you,” the father, to fight for them in the war in heaven. The daughter continued in the world as if the two years was all you had, so you got the ending for whatever her stats were at that point.

What’s the testing process like for games that have extensive scripts? How long does it take to make sure everything was accurately translated, looks good in the final project and, most importantly, works?

TT: It really depends on the type of game. If your game is a fairly straightforward branching game, it is easy to make a test plan that hits all of the branches. In Princess Maker 2, we were all new at this, and so mostly we just played the game a bunch. It wasn’t actually finished until I had already left for Working Designs, so I kind of trusted that everybody didn’t screw it up from the original Japanese version. I remember that Nebel wanted to add some extra code that changed combat responses to match a game that we all liked, but I vetoed it because nobody but us would get it and it could introduce new bugs.

Since the game was half in assembly language, there was always a chance that new stuff in the PM-BASIC code could leave a register in a strange state that could screw things up down the line, so we tried to leave the PM-BASIC in original condition as much as we could. Princess Maker 2 is a bit of a challenge, because it is really a simulation game, and so you have to actually be able to manage all of the math to get all of the branches.

How does it feel to see the game finally out there, but with what most would agree is a rough translation?

TT: It is rather bittersweet. We always wanted people to see and play the game, that is what motivated us. But the folks translating this version either didn’t know English well, or didn’t care enough to do a good job. They could have had our translation cheap, so there is really no excuse for that. It isn’t like we could sell it to anybody else.

How was it working for Working Designs? It was certainly no stranger to business issues during localization, but the company is still revered today for its work in that era.

TT: It was nice working for a small company and really being able to make a difference in the way the game came out. I did some tremendous work there that I am very proud of, especially my work on Magic Knight Rayearth! It was kind of a family business though, and sometimes family issues would come up that wouldn’t happen at a larger company.

What roles did you take on when working on Magic Knight Rayearth?

TT: I was the only programmer. I also handled things like art changes, integrating new video, some light translation, and for a brief time I did some of the rewriting. Victor [Ireland] didn’t like my rewriting because I was too close to the source material. I didn’t really rewrite for humor.

Reprogramming included writing a proportionally spaced text engine, shoehorning English text into their existing engine, adding more save slots to the memory card, replacing audio and video in already compressed cinepak movies without uncompressing them, writing command line tools to handle art modifications and hacking the game to find the art to modify, and many, many other things.

How have things been recently and what have you been working on?

TT: My last two projects were Rhythm Core Alpha and Rhythm Core Alpha 2 for Nintendo DSi and Nintendo 3DS. It is a music creation system that I have completely written on my own, and it is distributed through the Nintendo eShop as a download. Please go and download it now. I’ll wait.

Currently I am working on a hardware music synthesizer keyboard called the NanoEgg that gives you advanced modulation capabilities in a virtual analog synthesizer with a nice wooden case and a mini keyboard! You can see my progress on Hackaday.io. I decided that I was tired of working on other companies hardware and being subject to their whims, so I would make my own!

Thanks for taking the time talk with us, Tim!

TT: Thanks for giving me the opportunity to talk about my fine projects! It is nice to know that somebody still cares about Princess Maker 2 after all of these years. I think it is still the best of that series.

This interview has been lightly edited for clarity. Thanks again to Tim for taking the time to chat, and to you for reading!