backsolve... --> class()es for special matrices ?

Martin Maechler Martin Maechler <>
Fri, 22 Jan 1999 11:22:06 +0100

>>>>> "JonR" == Jonathan Rougier <> writes:

    JonR> I attach my versions of solve methods for triangular and variance
    JonR> matrices.  The triangular method is a wrapper for backsolve()
    JonR> which allows both a trans flag and a right flag (ie giving the
    JonR> four different solves that you alluded to).  The variance method
    JonR> uses a cholesky decomposition and then two triangular solves (one
    JonR> with the trans flag set), as recommended by Golub and van Loan.
    JonR> I have taken care with the returning vector/matrix, as in the
    JonR> backsolve() patch!

    JonR> I was not in on the previous Matrix class discussion.  I use them
    JonR> in Splus, but from a purely aesthetic point of view I don't like
    JonR> the idea that R would need a Matrix add-on.  I think I would
    JonR> rather see the appropriate methods built into the base.

I agree [at least as long as we only do small wrappers, etc..].
My point was that I want to use the same class() names as S+'s Matrix
library -- in those cases where do (+/-) the same.


r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: