There is a very useful identity, which is,
e(a+b) = ea * eb
However this only applies when 'a' and 'b' commute, so it applies when a or b is a scalar for instance.
The more general case where 'a' and 'b' don't necessarily commute is given by:
ec = ea * eb
where:
c = a + b + a×b + 1/3(a×(a×b)+b×(b×a)) + ...
where:
- × = vector cross product
This is a series known as the Baker-Campbell-Hausdorff formula.
This shows that if when a and b become close to becoming parallel then a×b approaches zero and c approaches a + b so the rotation algebra approaches vector algebra.
Derivation
We will use the formula for an infinite series of an exponent:
e(q) = |
|
|
So multiplying two of these together we get:
e(a) *e(b) = |
|
|
|
|
Now we want to move the second summation outside the first part, but since the first part has already been summed 'n' times we need to substitute n with n-m to correct for that:
e(a) *e(b) = |
|
|
|
|
I don't think what we have done so far has implied any commutation because the operands are still in the same order.