This is a note to:
Strange attractors and Lyapunov dimension.pdf

Some intuition for the Liapunov dimension:
In 3D an attractor has 3 Liapunov exponents:
L1 > L2 > L3.

#2) For an attracting fixed point all exponents are negative.
    Thus (sum_{n=1:d} L_n) is zero for d = 0.

#2) For an attracting limit cycle in 3d:
    Two negative: L2 and L3 [contracting directions]
    One zero:     L1
    (sum_{n=1:d} L_n) is zero for d = 1.

#3) For an attracting Torus, one exponent is negative and
    the other two are zero.
    (sum_{n=1:d} L_n) is zero for d <= 2.

Now thing of something like Rossler. Then
    one exponent is negative ["vertical direction"]
    one is zero [the "rotation" direction] and the other
    is positive.
    (sum_{n=1:d} L_n) is zero for some d > 2.
    The "d" is obtained by interpolation.
EOF