← Back to team overview

dolfin team mailing list archive

Re: Evaluating dolfin for use

 

On Fri, Oct 22, 2004 at 10:15:50AM -0700, Hugo Duncan wrote:

> > Is ublas actively maintained?
> 
> Yes, news://news.gmane.org/gmane.comp.lib.boost.ublas

Good.

> > 1. The package needs to pass the simple test
> >
> >     apt-cache search package
> 
> I believe boost as a whole is available through apt.

Indeed, it's included with the libboost-dev package. This makes uBLAS
a candidate. We still need someone to invest some time in doing an
evaluation of and (possibly) a transition to uBLAS. If anyone is
interested, please let me know.

> > DOLFIN currently
> > uses operator overloading in C++ to create a language for variational
> > forms. The problem with this approach is that the form is
> > "interpreted" run-time (creating objects) which makes the assembly
> > slower than optimal.
> 
> It is possible to make this a compile time approach. Boost.spirit
> does this to allow the user to write EBNF, and generates parsers
> for them at compile time (http://www.boost.org/libs/spirit/index.html)

Looks cool, but I found it so much easier to write the form compiler
in Python.

> This is beginning to sound like boost advocacy, which wasn't the point!

It's always nice to get some feedback and new suggestions.

/Anders



References