Need better presentation of error of correlated point vector residuals.

[ccrook]

Currently SNAP only provides a coordinate by coordinate residual for SINEX/GX data, which fails to represent the significance of the residuals correctly as doesn't account for covariance