sieve-dev [Roy Stogner] Re: [Libmesh-devel] LibMesh namespace?

[Matthew Knepley]

On Mon, Jun 21, 2010 at 9:32 PM, Jed Brown <jed at 59a2.org> wrote:

> On Mon, 21 Jun 2010 21:09:40 +0000, Matthew Knepley <knepley at gmail.com>
> wrote:
> > You can say that about any name.
>
> But "ALE" has no connection to "PETSc" or "Sieve" (admittedly KSP, SNES,
> TS are arbitrary too).  It's an obscure connection, and I don't even
> know what it stands for.  But already PETSc has a "PETSc" namespace, so
> it would make sense to me that ALE go inside the PETSc namespace (at
> least if you're going to tout how correctly namespaced your C++ is).


It is arbitrary, but as you point out so are the other names. The idea
behind not using
PETSc, at the time, was to make it possible to use apart from PETSc. This
now seems
unlikely.


> > We both do mesh stuff. I do not see why it is "more legitimate" for them
> to
> > do it.
>
> Because it's the name of their library and has been for about a decade.


The name is so common that this argument is incoherent. I am sure I can find
other
libraries with Mesh in them which are older.


> > I thought this was changed for PETSc 3. Now it should be
> > sieve/Mesh.hh.
>
> In -dev:
>
>  petscmesh.h:#include <Mesh.hh>
>  petscmesh.h:#include <CartesianSieve.hh>
>  petscmesh.h:#include <Distribution.hh>
>  petscmesh.h:#include <Generator.hh>
>

Yep, this should be changed.


> It also looks like there is a bunch of unnamespaced non-C stuff in
> petscmesh*.hh.


Just functions, not objects. These are templates for C functions, so I kept
the same names (thus there should be no collisions).

  Matt


>
> Jed

-- 
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their experiments
lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/sieve-dev/attachments/20100621/dd979c91/attachment.htm>