Yo! Frankie Interview

I spoke with Chris Plush and Campbell Barton, two long haired, clever fellows working on the development of the open source game Yo! Frankie. Based on the open source movie (see a theme developing?) Big Buck Bunny, players will follow the evil Frankie “unleashed in a natural world teeming with innocent wildlife, all subject to his whimsical bullying”. Great, eh?

Tell us a bit about yourself.

Chris Plush: The Apricot project was an assembly of extremely different people from all across the globe coexisting for six months in one room with a single goal. The truth, though, is that there were many different goals among us throughout the project, even ridiculous ones, and even from the very first day. These included closing the source and going commercial with the game, straying from the Big Buck Bunny premise, directions in software development, and even the final game engine the product would be presented on. It was an incredibly diverse team, geographically and idealistically. How did we manage the differences? Time, common sense, and the amount of perseverance necessary weeded out any weakness in the integrity of the team and our goals, and ultimately we settled on a course of action that lead to worthwhile results we could be proud of.

Tell us a bit about Yo! Frankie.

Chris Plush: Yo Frankie! sees the devilish natured flying squirrel Frankie unleashed in a natural world teeming with innocent wildlife, all subject to his whimsical bullying. The game is a 3D platformer where the aim is bullying other animals and exploration of worlds.

Why did you base Yo! Frankie on Big Buck Bunny and not something completely original?

Campbell Barton: This premise was chosen before the Apricot team were selected since we knew Big Buck Bunny had a strong theme, avoiding time spent on character design and concept development.

As it happened some of the team members really didn’t want to go with the Big Buck Bunny theme, but those game designs didn’t have enough appeal to convince Ton to follow through with them.

Chris Plush: Not basing the the Apricot project on Big Buck Bunny was seriously considered the first few weeks of the project, but in theory the benefits of sticking to the movie premise outweighed those of striving for something totally different. By sticking close to the movie, we had assets, character designs, models, and textures, and the help of the Big Buck Bunny team for the few months in which our project durations overlapped. While the overlap didn’t prove its benefits to be outstandingly worthwhile, the assets created by the Big Buck Bunny team proved much more useful than originally thought. The designs, models, and textures saved a lot of effort.

Why did you choose the nefarious butterfly killer Frankie to be your main character?

Campbell Barton: Sacha from the Peach project had a hand in this - Big Buck Bunny is a movie about getting back at bullies.

Playing one of the bullies was an obvious choice because they have more interesting actions and a simple motive. We did seriously explore playing the Buck the rabbit - but c’mon! Frankie the flying squirrel is much more bad-ass.

Chris Plush: The main character Frankie was not originally chosen to be the star in this game. It was the bunny that was initially queued up for the leading role but after considering the popularity Frankie received from Big Buck Bunny, and after examining the benefits of a character that isn’t sloth-like, has gliding abilities, agility, and a very well developed personality, we agreed Frankie would be the better choice.

Why an open source game?

Campbell Barton: The Blender Institute had just started out as a studio to work on open projects.

Inspired by the success of the open movie “Elephants Dream”, we knew open projects were a great way to establish a workflow users would benefit from, supporting in return by purchasing the DVD.

If we made a commercial game there would not be so much to set us apart from many other game studios, and realistically it would be hard to compete at this level with a small team in a short time.

Offering an open game with free assets and documentation and showing the process we used is much more unique and interesting.

What is your favoured OS to work on and also to support?

Our studio ran entirely 64bit-Linux workstations which was well suited to an environment where we were developing software as we used it.

Personally - I’m a big fan of linux since it gives you a lot of control and doesn’t try to impress me or be overly clever with how it manages resources.

Supporting any OS is a pretty tedious job, however I like the fact that I could be more involved with an open source desktop, reporting bugs and giving feedback to developers.

What’s Blender and CrystalSpace like to work with?

Both were undergoing development during the project so an artists life wasn’t exactly easy - they were guinea-pigs for features that were still in development, working around bugs where possible.

On the other hand they got to make decisions about what was important and influence the development direction somewhat.

Blender had the feature set needed to make game models from the start so we made safer, incremental improvements, whereas the crystalspace team chose to implement a new animation system, advanced shadowmaps, shader cache and a new mesh format. Possibly this was a contributing factor to CrystalSpace being unstable to the point the artists didn’t want to use it.

This is a tricky topic since the artists also knew blender better and there were more people at the institute who understood blender and how to manage it.

What were the challenges of working in an international team?

Campbell Barton: Just communication, meetings often went on longer then they might have otherwise.

You say on the site all you know about the characters are that they are funny and furry. How have things moved along since then?

I don’t think the characters were really further developed in the game since it’s just a jump’n run platformer. Though if you let Frankie idle for a while you’ll see his ego in action and get, well… out of control.

Any major changes during the project?

Chris Plush: This project has seen many changes. There were major changes almost on a weekly basis! Struggling at the start of the project to find common grounds among the team members and a work flow that produced promising results, to stay on schedule the only thing to do was scale down, and we were masters of the scale-down. So, to start, there were major changes in the game concept and planning.

It wasn’t easy going and within two months we had lost our first team member to the pressure, and he wouldn’t be the last rat to leave a sinking ship. However, because of one less particular body on board we bobbed back up to the surface and sailing from then on was relatively smooth. A character artist, Pablo, was brought aboard and things were kind of balancing out.

Then there were software issues, plus developer issues that vindicated the reality of these software issues. Although along with crystal space, B2CS (Blender2CrystalSpace), the plugin that allows integration of crystal space into blender, received a great amount of development that excited a lot of people, they weren’t able to achieve production level fast enough to ensure a desirable outcome. After some discussion it was decided the best idea would be for the game to branch off and use another game engine as well, the game engine native to blender itself. So there were essentially two teams, and two games being made, with the idea that the game being made in the blender game engine was a prototype that would then be duplicated in crystal space. Then the B2CS developer quit the team. The blender game engine ultimately became the delivery platform for the final game, and saw an immense amount of software development in both graphics, and game logic. These were developments no one expected from this project, but everyone was purely ecstatic about.

Would you do it all over again?

Campbell Barton: Definitely, though I’d try to focus on a smaller target and not thrash about so much. Making games is totally fun - even though many of us are not gamers, the creative and technical challenges are very interesting.

It would also be great to work on another game with the Blender Game Engine from the start, taking advantage of some features that were added very late in the project.

Anything you’d like to talk about?

Campbell Barton: Yo Frankie (blender version) is especially exciting because it opens up a lot of potential for user contributed levels, where any blender users can model a level and test instantly without having to use special level editing tools or dealing with converters/exporters.

But we really don’t know how the online community will respond. It is my hope that Yo-Frankie will at least be maintained at its current working state. I’d like to see extra levels to make Yo-Frankie a complete game.

On the other hand, it will be very challenging to manage contributions that deal with character logic, AI, adding new actions or characters that may not fit with the theme, or be intended for a different style of game.

Whatever happens, we will have an online subversion repository for continued improvements, I’m looking forward to see what users come up with in the way of their own games, building on what we have done.

I’d also like to mention how much we appreciate the animations from Moraes Junior ‘mangojambo’ (Brazil) who added some of the funniest actions for Frankie and single handedly created an entire character (Momo the monkey), so Frankie has someone to play with (that’s right two player split screen!).