Representing motion in 3D
To model motion of a body, we can use a transform (T) to transform each point relative to the body to a position in world coordinates.
If we want to model both liniear and angular (rotational) motion then we need to use a 4x4 matrix to represent the transform.
If we only need to model rotation then, we can reduce this to either use a 3x3 matrix or a quaternion.
Modelling rotations in 3D is not always intuative. For instance, a rotation of 90 degrees about the xaxis followed by a rotation of 90 degrees about the yaxis is equivilant to a rotation of 90 degrees about the zaxis. More information about rotations.
Frameworks, Chains (or Skeletons)
A lot of mechanical objects in the real world consist of solid sections connected by joints. Examples of this might be:
 Creatures such as humans and animals.
 Car Suspension
 Robot arms
 Ropes, string and Chains.
 Etc.
Where you have something like a robot arm, each section of the arm can pivot, bend and/or rotate from the previous stage. These sections and joints are known as a 'chain', if we are talking about creatures then it could be referred to as a skeleton. Similarly in biological terms a movable section can be referred to as a bone, and the attachments between each bones as joints.
Motions of chains can be specified in terms of translations and rotations.
 Forward Kinematics  From the amounts of rotation and bending of each joint in an arm, for example, the position of the hand can be calculated.
 Inverse Kinematics  If the hand is moved, the rotation
and bending of the arm is calculated, in accordance with the length and joint
properties of each section of the arm.
Forward Kinematics
This Chain is represented here, without any information about its actual shape, what we are interested in is the relative positions of all the joints. The first link is the base (or "anchor point" in nautical terms) is the end of the chain which always has a fixed or known position. From there, each link can be rotated (possibly in different dimensions depending on the degree of freedom of the joint). The other important factor is the distance between the joints, this can be represented as a translation in a TransformGroup. This transform group can also hold the rotation of that link. Therefore this chain can be represented by a hierarchy of TransformGroups. 
So Forward Kinematics allows one to find the position of the endpoint (and all the points inbetween) relative to the base. This is easy to do, just multiply by the TransformGroup matrixes in turn.
Inverse Kinematics
For inverse kinematics (IK), the position of the end point is known, and we need to find the angles of the joints. This is a much harder problem, there may be many possible answers, or there may not be a set of angles that would reach to that point. Although this is hard to do, it is useful, for example, when posing a model it is useful to be able to drag the extremities with a mouse and have the arms/links position themselves appropriately.
There are two approaches to solving inverse kinematics:
 Analytical  requires a lot of trigonometry or matrix algebra
 Iterative  better if there are lots of links and degrees of freedom.
IK  Analytical approach
If there are only two or three links then it may be possible to solve it analytically. One possibly might be to draw out the arm with the angles shown on it, then solve for the angles using geometry, a good knowledge of trig identities would help here. The problem is that this is not really very general approach.
Another analytical approach is to represent each links rotation and translation by a matrix. The end point is then given by all these matrixes multiplied together, so we just need to solve this matrix equation. Then find what rotation each matrix represents.
There may be many solutions or there may not be any solutions. In other words there are lots of ways to reach to a given point, or it may be out of reach.
If there are many solutions, then you might need to apply additional constraints. For instance, human joints can only bend within certain limits.
IK  Iterative approach
This is a more general approach for programming complex chains.
Start off with the joints in any position, then move each of the joints in turn, so that each movement takes the endpoint toward the target. Starting with the joint nearest the end point, rotate the joint so that the current end point moves toward the required end point. Then do the same with the next joint toward the base and so on until the base is rotated. Then keep repeating this, until the end point is close enough to the required end point or if further iterations are not moving it closer to the required point. It may be possible to have a more realistic strategy than this, for instance, if I am using my arm to pick up an object then, if the object is a long way away, I will move the bigger joints in the arm, then as the hand gets closer the smaller joints of the hand are used for the fine adjustments. The angle of rotation for each joint is found by taking the dot product of the vectors from the joint to the current point and from the joint to the desired end point. Then taking the arcsin of this dot product. To find the sign of this angle (ie which direction to turn), take the cross product of these vectors and checking the sign of the Z element of the vector.

Human Animation
A human skeleton can be represented as a hierarchical set of joints, see following diagram from mpeg4This can be represented by a scenegraph tree of transform groups (see this extract from VRML hanim spec, just a small section of human skeleton)
HumanoidRoot : sacrum sacroiliac : pelvis  l_hip : l_thigh  l_knee : l_calf  l_ankle : l_hindfoot  l_subtalar : l_midproximal  l_midtarsal : l_middistal  l_metatarsal : l_forefoot  r_hip : r_thigh  r_knee : r_calf  r_ankle : r_hindfoot  r_subtalar : r_midproximal  r_midtarsal : r_middistal  r_metatarsal : r_forefoot vl5 : l5 vl4 : l4 vl3 : l3 vl2 : l2 vl1 : l1 vt12 : t12 <snip>...
This skeleton is then combined with a physical
model of the human/creature to produce the animation