Re: [boost] Re: Standard Matrix interfaces

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2001-08-21 10:57:48

----- Original Message -----
From: "Hayati Ayguen" <Ayguen_at_[hidden]>

> There are many algorithms and algorithm interfaces implemented in itl
> which builds on mtl-2.x. even the multiply-function of mtl-2.x
> implemets/specializes to many algorithms.
>
> the flaw of mtl-2.x was that some algorithm-interfaces weren't
> specialized or even implemented: i believe to remember one of this was
> building submatrices from sparse matrices. there were some other
> problems but i can't remember.

Yes, MTL-2 has quite a number of problems. Many of them are soluble and not
surprising considering the evolution the state-of-the-art in C++ and generic
programming since it was written. Others are more closely tied to its
architecture. Still, seen as a research project, MTL-2 uncovered many useful
distinctions and concepts which are going to be important in whatever the
next thing is.

> before making unnecessary discussions:
>
> how far ist mtl-3? is it far enough to look at the source for discussing
> its interface wether special issues are considered?

Please take a look at the boost cvs repository on the matrix_development
branch:

cvs -d:pserver:anonymous_at_[hidden]:/cvsroot/boost login
cvs -z9 -d:pserver:anonymous_at_[hidden]:/cvsroot/boost
co -rmatrix_development -dboost_root boost

Or visit:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/boost/boost/?only_with_tag=ma
trix_development#dirlist

And browse the libs/numeric and boost/numeric subdirectories.