Drag/touch move input

[lgarron]

In principle (and as suggested by three.js demos), it's not hard to capture mouse/touch events and use them to animate moves on the cube. I think the main issue here is handling the in-between animation while someone is moving a layer.

Should include:

• rendering the (partial) move, with live updating
• good mobile support
  • needs to interact with touch support from the touch branch.

[mikebolt]

I am going to work on adding this feature. Here's how it works in the Cuber project (and now ThreeTwist):

1. Calculate the cube's bounding box within the cube's coordinate system, and save it as a Box3.
2. Use the cube's .matrixWorld property and the current camera to somehow create a "mouse ray". I'm not completely sure how this works but I know that the code for this is in Cuber's projector.js file, so I can reference that.
3. Find the intersection point of the cube and the ray.
4. Using that intersection point, figure out which face was clicked. Save a reference to that face's THREE.Plane.
5. Set some kind of flag such as "interacting" to true, until the mouse is released.
6. If the mouse moved, find its intersection on the face's THREE.Plane using a technique similar to step 2.
7. Calculate the direction vector, which is the difference between the current ray/plane intersection and the initial intersection point.
8. Take the cross product of the direction and the face's axis (the THREE.Plane's .normal property). This is the approximate axis of rotation.
9. "Snap" the axis of rotation to one of the cube's six axes. For the cube, I just did this by finding the axis with the closest axis vector. There may be more efficient methods.
10. Calculate the "sliding axis" using a cross product. It is the axis that is parallel to the plane, perpendicular to the rotation axis, and points in the same direction as the direction vector.
11. The angle of rotation is proportional to the magnitude of the direction vector projected onto the sliding axis. This can be calculated using a dot product. There may be other ways to calculate an "angle", but this one works well.

[lgarron]

I am going to work on adding this feature.

You're welcome to try. :-)

However, the code is very old and badly designed. I'm hoping to restructure it at a hackathon some time (the core part was written at a hackathon 5 years ago(!)), but I guess it doesn't hurt to add yet. Thanks for your description of how to handler it. :-)