Scaling for orthogonality

In updating algorithms where orthogonal transformations are accumulated, it is important to preserve the orthogonality of the product in the presence of rounding error. Moonen et al. (ibid., vol.39, p.1911-13, 1991) have pointed out that simply normalizing the columns of the product tends to preserve orthogonality-though not, as DeGroat (ibid., vol.39, p.1913-14, 1991) points out, to working precision. The authors discuss the previous work and give an analysis of the phenomenon.<
>