Re: [boost] Re: uBLAS formal review

From: Joerg Walter (jhr.walter_at_[hidden])
Date: 2002-07-04 01:52:32

----- Original Message -----
From: "Benedikt Weber" <weber_at_[hidden]>
Newsgroups: gmane.comp.lib.boost.devel
To: <boost_at_[hidden]>
Sent: Wednesday, July 03, 2002 1:59 PM
Subject: [boost] Re: uBLAS formal review

> "Joerg Walter" <jhr.walter_at_[hidden]> wrote
>
>
> > > 1) A member function that gives a pointer for the raw data in stead of
> > > having to write & Matrix(0,0) for IO and interfacing to FORTRAN
> routines.
> >
> > These member functions have been added recently and should be documented
> at
> > least in the 'Interface' section. We'll document them fully later.
>
> We also need an explicit statement in the documentation about the internal
> storage layout.

Agreed.

> I guess it is comformant to BLAS for triangular and banded
> matrices.

Yes.

> However I saw some discussions on the internal representation of
> Hermitean matrices.

BLAS conformant, too.

Regards

Joerg