Discussion: new operators for numerical computation

[John Lull]

"Pete Shinners" <pete at visionart.com> wrote (with possible deletions):

> but while we're at it, i would hope that we don't overlook the
> potential of elementwise assignment operators. sometimes those
> matrices can get rather large, and rather than create a new
> temporary matrix for every operation, it would be nice to have
> it done "in place"
> 
> MatrixA (*)= MatrixB
> MatrixA (.)= MatrixB

For all the other augmented assignment operators,
   A op= B
is precisely equivalent to
   A = A op B

If * is elementwise multiplication then *= should naturally be the
corresponsing elementwise operator, and
   A (*)= B
should be precisely equivalent to
   A = A (*) B

Anything else means these operators are different from standard
operators -- and that leads to confusion.

Regards,
John