Im a little late to this but I would like throw in a little.
First I really like the idea of a small accelerometer board with the same format as the blink-m. Using the sparkfun example there is absolutely no reason why these couldnt be produced for around $20 or so depending on whether the levels need to be shifted or not and what extra hardware is needed. (I also agree hacking the nunchuck is just in-elegant and wasteful.) The surface mount is simply not a problem when you manufacture boards yourself using the famed laser stencil/skillet method which is how every single board sparkfun sells in made. This method is simple and fast for reasonable sized batches of boards. Ive been too insanely occupied recently but I hope to introduce a new project that will use this fabrication method within the next couple of months.
Anyway, I am very much the hardware guy and if you need someone to manufacture something like this I might be able to help. But what would really make something like this shine is the supporting libraries also like the blink-m. And here I think is also a rub with the open hardware movt... not only do you have to get a bunch of folks to agree on the same tenants of hardware design, there's got to be those folks who will also get the software to work too. So, if there is enough interest in a 3axis uShield maybe that should be spun off as another topic for discussion and collaboration.
As I was thinking about this, a never ending string of stacking uShields could get a little tricky amp wise. Please correct me if Im wrong, but the blink-m uses two of the arduino (analog) pins as source for 5v and ground. This will only be good for 40mA correct? Stacking them is an open invite for some dead i/o pins if you ask me. Much beter to have the convenience of sticking on a little piece of kit that gives some functionality (such as a bad ass blinky light OR 3axis sensor, etc) and when you need the kitchen sink you can whip up the breadboard with the appropriate wiring. The trade off of convenience versus full functionality rather than trying to be the be all end all.
Finally, I would just like to say that I ABSOLUTELY love the chaos that happens with the Free/Arduino development. Its like a scene from the Smurfs, where the sun is out, the grass is green, and the butterflies lazily fly by when you are first introduced to this cute little blue guy that promises fully web enabled wine racks or blenders that speak to their owners. Then you pull back the curtains just a little bit and theres all this stuff being made, developed, proposed, and so on. Its like going to your first punk show in the 80s and having the biggest WTF moment of your life. Tod had been talking about 'smart leds' for a couple years and then one day out of the blue (to most of us) the happy arduino has a new blinky friend. Yeah its of lot of work and we all have our reasons for doing it but this chaotic development is downright fun if you ask me. Sure there could be more coalescing amongst some of us as we develop new things but I am very happy with this convoluted, prismatic, and even sometimes problematic landscape that has sprouted up around this platform. And that to me is an advantage of the open hardware movement and I applaud the foundation for their work and foresight to allow that to happen as well as the indie makers that in part made things happen anyway.
Cheers,
Brian