This applet will eventually have a picture in it's background, and you'll need to position/place/stretch the grid to match certain parts of the picture. Floors, walls in certain situations. Maybe mirrors, probably the surface of water.

Also, check out newest applet. This isn't much yet, but eventually it will be an entire little stick figure. It's only doing inverse kinematics in 2 dimensions currently. I'm having trouble figuring out how to apply the math to 3 dimensions.

Instructions: Click and drag the green circle.

Anyway, one of the benefits of specifying the ground or floor in a photo will be to let the code help you position the stick figure. If we know where the floor is, then a nifty little javascript can help to ensure that people aren't floating around, or have their feet embedded in the floor. We also get to show how much of that person is under water.

On the stick figure, I've been thinking of a few ways to make it easier to use. I've always intended some basic interface that mimiced the important parts of Poser. For those unfamiliar with it, to the right of the human 3d model there were a set of dials, to allow you not only to fine tune the rotation or movement of whichever bodypart was selected, but also it's size. Still, it's going to take some skill to use, even if it's as easy as I hope.

So, I was also considering a dropdown list, with the names of as many as 20 or 30 base poses. Standing, sitting, kneeling, etc. Selecting any of these would position the stick figure in that pose instantly. Certainly doing this and making minor adjustments to that pose would be easier than trying to go from the "standing with arms outstretchde" pose every single time.

Still harder to decide yet, is whether an actual stick figure is going to be adequate. It's hard to tell which way it's facing with no depth cue. Also, there are some things that can only be done if I render an actual 3d mesh. It may be possible to fake more of a solid feel to it, without actual rendering. For instance, why draw a single line from hip to knee, or render out a few dozen triangles for that, when you could just draw an approximation of the thigh's outline with a freeform polygon (unshaded of course)? Still, how figure the silhouette without doing the actual rendering? Anyone have any insight on this?

Finally, anyone with any pointers or links to data on the average size/lengths of adult humans various proportions would be welcome. At some point, I'll need to know the length thigh and shin bones for people of various heights, and I don't think a tape measure and a full-length mirror will be much help.