1. It is programming. Kodu's a real, albeit small and specialized language. Simple things are very simple to do, and complex things are possible. If you want to do something super-complex, you will have a learning curve, but nothing like what you'd see with a conventional programming language.
2. It can't do absolutely everything. We managed to squeeze in just enough (mainly camera support) for side-scrolling games, but there are still little annoyances: for example, you can make roads float in the air, (such as for a jumping game) but if two roads cross over each other, you can only place objects on the higher road. Annoying; we just didn't get to it.
3. The characters all move differently to support different design goals. If you're into fast, twitchy action, go with the saucer, wisp, or puck, all of which can turn on a dime and accelerate very quickly. Most of the other characters are a little slower, some a lot.
4. Everything has a cost. Sometimes some fairly simple things can be kind of expensive when the game is running. We try to warn you when you've got to much going on in your world - look for the thermometer - but we err on the side of letting you go for it, so it's up to you to keep things running smooth by trading off how you're using different features. For example, if you have a bunch of characters that are all trying to look at each other (using the "see" sensor,) they'll be doing a lot of expensive tests against the terrain. This applies to other things as well: for example, if you drop 50 coins in a level and program them all to react when bumped, they're all doing a little bit of thinking each frame that really adds up. Better to program the character to detect the coins, so you only have one brain running rather than 50. Hint: hearing is more efficient than seeing, because it doesn't have to check if something's blocking.
5. It's a 1.0 from a small team. We do work at Microsoft, but the Kodu team (design, dev, and test) is only six people. I'm sure we missed something. We are standing by to fix any bugs the minute they appear and to flip a service build quickly if necessary. I've been in software long enough to be quite sure we'll need a refresh at some point.
6. If you've been in the playtest or review, your worlds will not be available in the retail version. This is a security thing on the Xbox 360; we can't do anything about it. If it helps, I've rebuilt dozens of levels many times. You get pretty fast at it.
7. The built-in games are just a start. We've put a good double handful of prebuilt content, all built by the team and our early testers. We expect you can do way better. We have designed each of these worlds to show some realistic techniques. We'll be doing some deconstructing of these levels on the blog so you can see why we "did it like that." It's our hope that you can find the world that is closest to what you're thinking of, and then go from there.
8. Kodu is for making small games. We considered many features that would support very large worlds and very long campaigns, but were very conscious that the toolset stay simple and streamlined. Some of these calls were wrong, but we are very happy with the balance we came up with - a set of quick, simple tools for making very cool small games. If you're careful with performance, you can make significantly larger worlds, but you're not going to make Gears 3 with Kodu 1. Haiku is a word that comes to mind. Here again the built-in levels show you some of the tradeoffs you can make.
9. Use the tweak screens. Select a character and press
10. You are Kodu. The success of Kodu depends entirely upon what people build with it. I have been really surprised by what people have pulled off with it already, and supremely delighted to see the level of buzz in the community. It's out of our hands now, and we are counting on all of you to realize the dream of the Kodu omniverse. We're all just getting started...