Some errata from Alexander Gouberman.

LaTeX notation is used.

- p. 6: "Jordon" -> "Jordan"
- p. 10: Proof 1.4.2: p(A) E_i = \sum ... * (A-\lambda_i) E ... -> the expression (A-\lambda_i) must be (A-\lambda_i)^k
- p. 12, above Eq. (4): m_i -> m_1 in r(\lambda).
- p. 34: close to the end: in B\hat_\lambda = ... = I - \lambda A\hat" -> must be "A\hat_\lambda"
- p. 35, top: must be "T\inv A\hat_\lambda T" instead of "T\inv A T" (?), etc.
- p. 39, 2nd paragraph: "But then ... x(t_0) for all t" -> must be x(t)
- p. 81, top: compute -> commute
- p. 81, mid page: "Hence taking ..." in (I - \lambda\inv A\hat B\hat^D)\inv the "-" must be "+"
- p. 88: Theorem 5.2.1: limit does not hold for t=0 (case B = 0 is considered special in the proof) at p. 97 below X_0(0) = I - BB^D assumes that the limit holds at t=0, i.e. X(t) is continuous
- p. 90, bottom: "r_0 = ||A_12|| + ||A_22|| + 1" -> A_12 must be A_21
- p. 92: "Thus F_0(\eps) = 1/(2pi*i) \int (\zeta - B)\inv ..." -> in the integral there is a \zeta factor missing