fenics team mailing list archive

["Johan Hoffman" <jhoffman@xxxxxxxxxx>]

> On Tue, Sep 16, 2008 at 10:46:11PM +0200, Johan Jansson wrote:
>> Anders Logg wrote:
>> > On Mon, Sep 15, 2008 at 11:14:50PM +0200, Johan Hoffman wrote:
>> >
>> >> Yes, this is natural. I am all in favour of including Unicorn among
>> the
>> >> fenics-apps, as you describe.
>> >>
>> >
>> > The only difference would then be that the repository for Unicorn is
>> > under
>> >
>> >   fenics.org/hg/unicorn
>> >
>> > instead of
>> >
>> >   fenics.org/apps/unicorn
>> >
>> >
>> >> But at the same time, I also think it is
>> >> important to keep Unicorn as a FEniCS project in itself, since it is
>> the
>> >> only project aiming at realizing the FEniCS vision in the direction
>> of
>> >> mathematical modeling (turbulent/laminar compressible/incompressible
>> flow,
>> >> fsi, etc. as one solver).
>> >>
>> >> /Johan
>> >>
>> >
>> > I don't agree. I also want to do all those things, but as part of
>> DOLFIN.
>> >
>> >
>> Hi!
>>
>> I'm very positive about FEniCS Apps, and I think that Unicorn should be
>> included as well.
>
> Good. Note that being part of FEniCS Apps is not the same thing as not
> being a project. Everything inq FEniCS Apps is a project. The point as
> I see it is to collect application-oriented projects (applications)
> and put up an infrastructure that allows us to add new projects at
> high speed with minimal resistance. If Unicorn is placed under FEniCS
> Apps, it would still be a project, still have a web page, still have
> its own repository etc, just like DOLFIN-grade2 and FEniCS Plasticity.

Good. But then I think that essentially we are all on the same page.

>
>> In light of this discussion, if you want to develop a solver for
>> compressible/incompressible flow, shouldn't that be included in FEniCS
>> Apps, or? Why would that fit into the DOLFIN project?
>>
>>   Johan
>
> Yes, it should be in FEniCS Apps (the solver) but as much as possible
> of common infrastructure (mesh algorithms etc) should be in DOLFIN.

Ok. Sounds very much like the Unicorn approach, which I think is good for
DOLFIN.

/Johan


>
> --
> Anders
> _______________________________________________
> FEniCS-dev mailing list
> FEniCS-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/fenics-dev
>