Re: GSOC 2013 - FEM LIBRARY

[c.]

On 30 May 2013, at 16:06, address@hidden wrote:

> Message: 1
> Date: Thu, 30 May 2013 08:56:07 -0500
> From: Max Brister <address@hidden>
> To: Marco Vassallo <address@hidden>
> Cc: "address@hidden" <address@hidden>
> Subject: Re: GSOC 2013 - FEM LIBRARY
> Message-ID:
>       <address@hidden>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> On Wed, May 29, 2013 at 6:01 PM, Marco Vassallo
> <address@hidden>wrote:
> 
>> Hi everyone,
>> my name is Marco and I'm working on the implementation of a Finite
>> Elelment Library.
>> The idea is to create wrappers in Octave for Dolfin, which is the
>> problem-solving
>> C++ environment of Fenics.
>> Something similar has been done for Python, but we want to go directly
>> from Octave
>> to Dolfin instead of using the Octave interpreter for Python.
>> So, first of all, I need to check the compatibility between Octave and
>> Dolfin.
>> I noted that, unfortunately, including in the same .h file both the Octave
>> and
>> Dolfin header file generates some error at compilation time because both
>> Octave and Dolfin define a class "Array". Does anyone encountered
>> a similar problem somewhere?
>> Thus, as a first step, we want to write an Octave function for reading
>> Dolfin mesh (.xml) and converting it into the Octave format. This function
>> could be added to the msh package.
>> I will try to keep you updated/discuss also on this blog
>> 
>> http://gedeone-gsoc.blogspot.it/
>> 
>> Marco V.
>> 
> 
> The proper solution to multiple definitions of "Array" is to use namespaces
> in either Dolfin or Octave (preferably both).


The idea of introducing an Octave namespace has come up many times,
but has never been done, does anyone recall why?

As for Dolfin, it does have its namespace 

Is that any reason not to do the same in Octave? 
I mean, apart from the boring work it takes to add the namespace
clause in all headers and sources in Ocatave and of 
Octave-Forge of course ...


c.