[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: midterm evaluation
From: |
c. |
Subject: |
Re: midterm evaluation |
Date: |
Tue, 30 Jul 2013 09:54:12 +0200 |
On 29 Jul 2013, at 19:50, c. <address@hidden> wrote:
> Marco,
>
> This is a reminder about the GSoC midterm evaluation deadline.
>
> The deadline for midterm evaluation is Friday August 2,
> could you please prepare a blog post outlining the current status
> of your project by Wednesday as we agreed during our meeting on Friday?
>
> Also, please remember that you have to rate your mentor as well.
> This is partiicularly important as not filling in the mentor review will
> result
> in failing the midterm evaluation, regardless of how well your work will be
> rated!
>
> Best wishes,
> c.
Hi Marco,
Thanks for your blog posts, the usage examples are very cool,
I think the project is in very good shape and it is progressing
at a good pace, there should be no problem in completing all
planned tasks in due time.
As for the midterm evaluation I think you shoud add another post
describing in more detail:
1) relevant implementation details of the currently available methods
2) main issues to be solved in the currently implemented methods,
in particular the issue with allocating sparse matrices and that
of registering new types in the interpreter.
3) Plans and timeline for the remaining part of the project.
Finally some comments on things to be improved:
1) I am Extremely unhappy with the naming conventions you are using,
I think it is vital to make Octave-FEnics scripts as similar
as possible to the equivalent Python/C++ counterparts.
This requires using the EXACT SAME names for any type or
function whenever possible. This means in particular the "fem_"
prefix MUST go.[*]
2) One of the great advantages of having a direct interface from
Octave to FEniCS is that it allows to use FEniCS features in the same
environment as other great Octave tools, it would be great if you could
demonstrate this in your examples, e.g.:
2.1) Show in an example how existing functions
in fpl can be used to export or visualize
computation results.
2.2) Show how ODE solvers in Octave can be used in conjunction with
dolfin to solve evolutionary problems.
You should also mention this latter suggestions in your plan for the rest of
the project.
Best wishes and keep up with your great work!
c.
[*] Personally I would also remove it from the package
name, even if others disagree. I don't think messing up
with package names is a good substitute for package
categories, or for having a serach feature in Octave-Forge/Agora.
The python module is named "dolfin", not "fem_dolfin" why should
we do differently?