[Changeset] Add the rcond function

[David Bateman]

Although this changeset is large, most of its just replacing "rcond"
with "rcon" to avoid shadow warnings for the rcond methods that the
changeset introduces in Matrix and ComplexMatrix classes. The oct-file
itself is then trivial.

Note that although I started doing the same methods for sparse matrices,
matlab doesn't support this and so I didn't do it in the end. Use
condest(a) if the matrix "a" is sparse.. Also it appears that matlab
only uses the LU factorization for their rcond function and not the
hermitian or triangular matrix versions of the LAPACK factorizations.
Therefore the Octave code will be faster for these cases, but won't
necessarily return the same result as matlab. This is a 1-norm estimate
of the reciprocal condition number in any case and so I believe the
difference can be ignore as its not the exact value we are interested by
rather its magnitude.

D.

--
David Bateman                                David.Bateman@...
Motorola Labs - Paris                        [phone] (Ph)
Parc Les Algorithmes, Commune de St Aubin    [phone] (Mob)
91193 Gif-Sur-Yvette FRANCE                  [phone] (Fax)

The information contained in this communication has been classified as:

[x] General Business Information
[ ] Motorola Internal Use Only
[ ] Motorola Confidential Proprietary

[John W. Eaton]

On 14-May-2008, David Bateman wrote:

| Although this changeset is large, most of its just replacing "rcond"
| with "rcon" to avoid shadow warnings for the rcond methods that the
| changeset introduces in Matrix and ComplexMatrix classes. The oct-file
| itself is then trivial.

I applied this changeset.

| Note that although I started doing the same methods for sparse matrices,
| matlab doesn't support this and so I didn't do it in the end. Use
| condest(a) if the matrix "a" is sparse.. Also it appears that matlab
| only uses the LU factorization for their rcond function and not the
| hermitian or triangular matrix versions of the LAPACK factorizations.
| Therefore the Octave code will be faster for these cases, but won't
| necessarily return the same result as matlab. This is a 1-norm estimate
| of the reciprocal condition number in any case and so I believe the
| difference can be ignore as its not the exact value we are interested by
| rather its magnitude.

OK.

Thanks,

jwe