Re: [boost] ublas: matrix clear() and operator=

From: Joerg Walter (jhr.walter_at_[hidden])
Date: 2002-05-03 15:18:27

----- Original Message -----
From: "Victor A. Wagner, Jr." <vawjr_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, May 03, 2002 6:11 PM
Subject: Re: [boost] ublas: matrix clear() and operator=

> At Wednesday 2002/05/01 12:17, you wrote:
> >----- Original Message -----
> >From: "Toon Knapen" <toon.knapen_at_[hidden]>
> >To: <boost_at_[hidden]>
> >Sent: Tuesday, April 30, 2002 6:09 PM
> >Subject: [boost] ublas: matrix clear() and operator=
> >
> >
> > > Currently, the matrices can be clear()-ed such that all elements are
set
> >to
> > > 0. I would find it even more convenient to be able to use
> > > operator=(some_value_type) to assign a specific value to all the
entries
> >in
> > > the matrix.
> >
> >How would you define the semantics of such an assignment operator for
sparse
> >matrices?
>
> It seems to me the whether the matrix is implemented as a 2D array or held
> in some sparse implementation, ALL of the entries should now have
> (some_value_type).

I just realized that my last answer to Toon's question missed the main
point: in ublas matrix containers are usually dense, packed (i.e. symmetric,
hermitean, triangular, banded) or sparse. If we use a scalar assignment
operator, we could assume a matrix layout. It seems reasonable to assume a
dense layout. The main point now is, that the proposed assignment operator
is rather useless, if one assigns a non zero scalar to a sparse matrix
assuming dense matrix layout.

IOW I'd like to keep scalars, vectors and matrices separate ;-)

> As an implementation detail for the sparse representation, there is
nothing
> sacred about calling the value of an unheld entry zero(0), it could be any
> value.  One possible implementation of this operator would to be to
"empty"
> the sparse cells which exist and set the "default value" appropriately.

Yes, in the zero case you're right.

Regards

Joerg