We Heart Katamari Preview

1up.com has up a look at the upcoming sequel to Katamari Damacy, set to launch in Japan sometime this summer. Besides some screenshots and discussion of new gameplay, the article discusses the original title and its place in the industry. From the article: "With intuitive, non-violent action and simple but stylized visuals, Katamari was one of those rare creations that managed to transcend the stereotypes and demographics of gaming and strike a chord with a wide spectrum of gamers, both casual and hardcore, male and female, youthful and elderly. The brainchild of Keita Takahashi, an artist turned game producer, Katamari's unique appeal can largely be chalked up to its creator's unique aspirations."

Comments

Re:I'm fed up of hearing about Katamari's sequel

[cgenman]

I read this and almost cried. Why don't game developers use the same design techniques and standards as other major software developers? Using hard-coded string values? Hard coding the software to run at 60 Hz? Pushing your memory budget so close you cannot fit a few more characters of text? Not only that, but game engines are re-written from scratch for virtually every new game; the wheel is re-invented 12 times a month in any given project.

A lot of it has to do with what you are trying to do with your budget. String lengths aren't checked because that would double the amount of milliseconds the system is spending on text. You can check all of that at export time, but that programmer time might be best spent on another aspect of development. You squeeze the system as hard as you can, and maybe everything including rendering, physics, etc takes 56-59 MS on average. If you reduce the amount of pixels per frame but increase the frame rate, the system may be pushing exactly the same number of pixels, but now you have to update physics 20% more often. Maybe you're using the sound processor to update a part of your physics model because you know *exactly* how much time it needs to spend on rendering your streaming audio, but now your timing is all shot. Maybe you've hardcoded string length because then you can allocate a single text array at the beginning of a level and shave a few cycles off of seek time.

It's not just shoddy code design, and I'm sorry if I made it sound like it was. The programmers I've worked with at game companies have been some of the best I've known or worked with anywhere. But with extreme optimization comes lots of dependencies. It's not a question of being a bad programmer, it's a question of what you're trying to optimize for. And yes, some of that comes down to optimizing for time and not having enough time on any given aspect. I've seen an entire AI system implemented, pretty well I might add, in 3 days by a single coder. You're given a budget, and you want to make the best game possible with that budget, which usually involves pushing things. But a lot of the localization related problems just come from having extremely optimized code and other resources.

And most engines aren't written from scratch, though few are bought. Pretty much every company has a 3D graphics, animation, and gameplay engine they've put through quite a few cycles, even if they're good about making it not look like that's the case.