“This is my tree. It makes music.”

It took “actual 12-year-old girl” (as Microsoft described her) Sparrow to rescue Microsoft’s drab CES keynote (and drab tech news week) and get us back into the Future again. That future is one in which the dazzling interactive 3D world of games becomes a playground you can shape. In this case, the showpiece is a game called Kodu, but that may just be the beginning. The reason all of this is so deeply significant is that what you need to make something work for kids could say a lot about how the rest of us will work with 3D interfaces, including for making art, performance, and expression.

From Channel 10, here are a couple of priceless demos of Sparrow using the tool. It seems scripted as first – and then you realize the “grups” are just trying to keep up with the kid. Sparrow, can you do the entire Microsoft keynote next year? I bet you’re already running Windows 7 beta anyway. (You’ll need Silverlight for the video – this is Microsoft — but it’s been running well for me these days.)


CES 2009: Watch Kodu in Action

This is the video from the keynote itself, which is even more frenetic:

Kodu, Getting 12-Year-Olds Making Xbox Games

Kodu (formerly Boku) is the latest in a long line of research into giving kids game design tools, and using games and interactive design to teach young people about programming concepts. There are several things about Kodu that make it special. First, there’s the distribution: Kodu goes straight to the Xbox 360 via the Community Games Channel. That makes it accessible on the device a lot of people are playing, and couples it with Microsoft’s own XNA development framework and community – something 12-year-old Sparrow might graduate to when she’s a high-school student. Second, it really looks like a game: the graphics are slick, interactions are fun and quick, everything is integrated with the Xbox gamepad and it just feels more like a game than an academic exercise. That’s not to say deeper tools aren’t valuable, but a casual tool could be a good way to first get people hooked. Third, the rule-based interface is really elegant. Sparrow blazes through applying behavioral attributes to elements of the game, making intelligent agents like the particle-based wisps. It’s not a new idea, but it’s great to see how nicely it works with the interface design and controller.

And most importantly, it’s important that Sparrow was on the keynote. That gives kids in general, and this kind of game design, new attention. Oddly, many bloggers simply dmissed Kodu by comparing it to the PS3’s LittleBigPlanet because it also happens to be a world-building environment. Luke Plunkett at Kotaku ran with the headline “Microsoft Inch Further Up The LittleBigPlanet Bandwagon With Kodu,” even though in the same story he admits it “looks nothing like LittleBigPlanet” and is less of a game and more of a “tool to help design games.” It’s hardly a “bandwagon” if the game is completely different, of course, and Plunkett ignores Kodu’s own research history.

But that’s how narrow press and public perception is when it comes to 3D interfaces. Games press allow countless identical shooters, but something that’s by-definition open-ended like LittleBigPlanet is determined to be a “novelty” – and no one’s interested in other takes on the same idea. The public, meanwhile, may assume games are just shooters and that games are just games, when in fact the same 3D interfaces can be learning tools or expressive tools or anything you can imagine.

Where the Research Comes From

Back to Kodu, you can tell a lot about what this is about just from the background of the Kodu team at Microsoft Research. Lili Cheng manages the creative and social groups at Microsoft and founded Microsoft’s Social Computing. She’s a veteran of Apple’s Advanced Technology Group, where she worked with user interface and QuickTime Conferencing and QuickTime VR. There, she would have presumably worked with ATG founder Joy Mountford, who also has a long history of advocating this kind of work; Joy joined me to talk about the future of what Joy calls “fluid” interfaces at a session I organized at last year’s South by Southwest.

Programmers are always interesting, too. Mark Finch has a physics background from Georgia Tech, a history at Cyan Worlds, and did those lovely visual effects in Rock Band with the folks at Harmonix. He’s not only worked on making programming friendly to non-programmers, but hardcore programmers, as well, a veteran of Emergent, a middleware maker.

Matthew MacLaurin, the program manager, describes Boku’s inspiration as (not surprisingly) Logo and HyperCard. As part of the generation that grew up on those tools, and now wishes I had spent more actual time with them as a kid, I can think of no better comparison.

This isn’t bandwagon work, folks. What’s interesting to me is the small group of people around the world who have been working on these kinds of problems for decades, and just how long it takes for those ideas to fully form and become marketable. You can pick a project and trace a designer or a set of influences back through time, and you have – well, a tree that makes music. In a futuristic industry that often lacks real imagination, you will find a common set of interests in tackling tougher, cooler problems, like making software more genuinely creative. The likes of LOGO and HyperCard are at the center of this; the fact that today, Macs and PCs don’t ship with anything that lives up to Bill Atkinson’s late-80s HyperCard, let alone surpasses it, is really unforgiveable.

You need to support that kind of work. Microsoft is still investing heavily in R&D as companies like Apple have vastly shrunk their R&D efforts. (The once-fruitful Advanced Technology Group at Apple is now long gone.) That is worth noting, especially as Microsoft also gives heavily to academic research. This isn’t a pro-Microsoft argument. But it is something I think can be generally valuable, and the fact that Microsoft R&D manages to do both far-ahead, future-thinking research and productize other developments quickly is not only commendable, but something some of its competitors could do more.

Other Tools, and Lots of Potential

My one big complaint about Microsoft Research is that so little of what they’re doing is open source. Microsoft is capable of embracing open source when it’s so inclined, and in this case, with Microsoft trying to push its own XNA development tools, it seems that open sourcing even a small part of the research would be a no-brainer. It doesn’t have to be open for the sake of it; there’s a clear reason for it. Having code available to XNA coders would benefit Microsoft, and teachers and developers may need the code to be open source to extend the educational and interactive possibilities of the framework. On the other hand, to build a tool that does those things, it may be necessary to make that part of t
he design goals, which isn’t necessarily Kodu’s goal. And there are other places to look.

Kodu aside, there are other tools working along these lines, and there’s clearly enough potential that in the near future we could see many, many of these tools, each filling different needs. For instance, Platinum Arts Sandbox is a real game engine. It’s open source. It’s easy for kids – and it’s also easy for adults. You can also edit cooperatively online, which seems to me an essential feature for these kinds of tools.

Platinum Arts Sandbox [Free + open source]

Whereas something like Kodu is limited to assembling stock sets of rather abstract interactions, Sandbox can cook up sidescrollers, RPGs, and the like. It looks incredibly powerful – you really do have friendly tools atop a game engine, which is something I think we’ll increasingly see even for developers, as gamers look for new and unique experiences that require more prototyping.

The Kodu folks even point at some tools that inspire them. One is Scratch, a Web-based game engine that, like Kodu, allows you to connect blocks to build interactions. (Tinker Toys or Legos for AI?) The interface looks just like Mark Coniglio’s early research project Interactor, which I thought was heavily underrated; Mark went on to make the popular visual tool Isadora, a big hit with the modern dance world.

Scratch @ MIT [Free]

You can go create a free account and start playing there right away.

There’s also the powerful, Java-powered Alice, developed by Carnegie Mellon researchers and supported by EA, Intel, DARPA (hey, they gave us the Internet, right?), the National Science Foundation, Sun, and others. Alice is a 3D programming environment with drag-and-drop environment that uses tiles to represent blocks of code – not unlike the way tools like Max/MSP/Jitter use visual patching to represent blocks of C, Java, and JavaScript and data flow and variables for that code. Alice looks terrific, and it runs on all desktop platforms. I’d love to see something like this that uses a newer engine and render pipeline, though – jMonkeyEngine, anyone?

Alice [Free]

But then, looking at these tools only makes you realize how much more could be done. Why aren’t there tools that make it easier for adults to learn game programming? (XNA takes some big leaps in that department, only to skip over other critical elements.) Why are there so few good prototyping tools for game designers? Why don’t we have visual and musical tools that work like this, and function like performance worlds instead of just game worlds? Why is collaboration and online community so often an afterthought?

There’s no time like the present to begin tackling these problems, because the tools necessary to develop tools that make it easier to develop worlds have gotten much more accessible. You have a fraction of the work that, say, Bill Atkinson had to do with HyperCard. And I believe, based on the people who are out there and the projects that are germinating, we really will see some massive tools in the future.

In the meantime, I’d love to do a roundup of tools like Kodu and Platinum Arts Sandbox, and other game design-for-kids options. Let us know if you have ideas.

But I can say, without question, Kodu was the most exciting announcement I saw this week from any tech company.

  • http://digitaltools.node3000.com Digital Tools

    I was just wondering then, how the Little Big Planet came along with their entireally new kind of concept. I think academic research may had been experimenting with that stuff a lot, but it needed at "blockbuster" just to switch the perception and to show, what the "next big thing" in games will be. And that game/tool was definitely Little Bit Planet. Microsoft hide their stuff in their labs, just waiting for the right moment to follow up the sophisticated "Bandwagon" – and people understand.

  • Pingback: Create Digital Music » Generative Music Interfaces of the Future – Look to Games?

  • HMC

    Hey! Sparrow did an awesome job with this! I look forward to introducing my kids to programming with Kodu soon. Looks fun and educational.

  • http://createdigitalmusic.com Peter Kirn

    @Digital Tools: well, I think "new" can be overrated. The thing is, you're seeing this now because the concepts — and how we approach gameplay, even as designers — has matured.

    If credit goes to anyone, I think it's Bill Atkinson (HyperCard) and Dan Winkler (accompanying HyperTalk). They really introduced the whole concept of using interactive objects to piece together hypermedia. And I seriously doubt that Little Big Planet is going to have the impact on the universe that HyperCard did, cool as it is. Remember the amount of content created with HyperCard, partly thanks to the fact that it was bundled with all Macs. It launched ideas about hypermedia that helped create the social change that enabled the Web. It was the engine in which the first version of Myst was built (further evidence that these things can become prototyping tools for new creations.)

    The emphasis in Kodu is especially on creating those kinds of interactive behaviors — it's just with graphical icons, although that in some ways makes it *less* expressive than HyperTalk.

    If LGP and Kodu have something in common, it's the desire to try to be an authoring tool and a game at once — which makes Spore something of the same ilk. I think it's impossible to judge at right this instant what will endure there, because you have to take a longer view.

    I also don't know that Microsoft kept this under wraps, so much as this is when they happened to finish it. But I'll try to get in touch with the development team.

  • Pingback: Create Digital Music » Microsoft Research’s Songsmith Will Sell for $30, Match Accompaniment to Your Singing

  • http://www.abstractmachine.net Douglas Edric Stanle

    Great article, as always.

    I happen to think (and I have been defending this idea for quite some time) that this and similar efforts are where we need to be when it comes to digital tools. Little Big Planet is really exciting, but mostly because of the quality and breadth of execution; whereas the idea itself is really quite old and can be traced back, as you mention, to Logo (and Lego), but also back to the Dynabook, Smalltalk, or Ivan Sutherland's work.

    In many ways, the computer's native "media" is its algorithm. There is no obscure ontological restriction obliging us to access these algorithms via code. We can use shapes, colors, sounds, objects, even subtle semiotic cues via body language (body/face/gesture capture, recognition — indeed: dance); basically anything that falls into the realm of the sign and probably beyond that.

    Personally, I use code because that's where all the powerful modularity resides (i.e. the creative part) and I don't have the time to build my own computer (all programmers collaborate, so we are dependent on each other's aesthetics and sensibilities). But there is nothing I do at the code level that couldn't be done in a form such as this. Code is an inherited constraint at best.

  • Pingback: Fun game design for kids | Honeste