Jim's note covers a classical method for producing a generalized inverse and one which is useful in programming multivariate statistical computations.
Richard Reyment -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
