Re: [Patch] fem-fenics patches

[Marco Vassallo]

On Wed, Mar 5, 2014 at 8:10 AM, c. <address@hidden> wrote:

On 4 Mar 2014, at 13:22, Marco Vassallo <address@hidden> wrote:


>> I've attached my modified source file so that one can read the code more
>> easily, though.  Hope this helps!
>
> Yes thanks it is easier now. I've just pushed the two patches you submitted.

I actually don't like these changes, or rather the phylosophy behind them.
I would suggest to back them up.

My idea was only that the patch submitted yesterday was an extension of what we already had.

The old syntax still works fine, and I wanted to add the matrix version as soon as possible using Daniel
contribution as a starting point.

There is good reasons why we want to stick to a syntax that is as close
as possible to other FEnics interfaces (either C++ or python).

Current FEnics API's  come with very extensive and detailed (and quite lengthy)
documentation, e.g.:

 http://fenicsproject.org/documentation/tutorial/index.html
 http://fenicsproject.org/book/index.html#book

Unless we want to repeat the tremendous effort required to produce those docs
we should really make any effort possible to stick to a compatible syntax
in order to make it as straightforward as possible to switch from python to Octave API
seamlessly.

Even if that costs  the effort of giving a little more thought on a change before
pushing it.

In this particular case, a change that would preserve the compatibility AND allow
the convenience of vectorization, is to allow the syntax:

 eval (FUNC, P)

where, if P is a matrix, evaluation is done by iterating through its columns.

This would also allow, for example, interpolating to the nodes of a new mesh
by simply doing:

 f = eval (FUNC, new_mesh.p);

> I wait for the problem with the configure.ac file to be fixed and then I prepare a new release.
> Thanks for your help.

I don't think we should , we are planning a new project to improve fem-fenics, rather than
rush into making a new release just now I think this is the time to sit and plan the future
package

If the patch provided from Juan Pablo and Eugenio fixes the problem I think it could be a
good idea to have a new release because almost every new user installing the pkg is running
in the same problems. But I agree with you that it is the moment to plan for the future development

of the pkg.

By the way, yesterday when I pushed the changes I made a mistake and I have created a new head
 in the mercurial repository. I think that I fixed it but I'm not sure. If someone could check that everything

is fine or point me to some reference which explain how to do it, it would be helpful.

Thanks

Marco

> Marco