[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Metapackage nanoscale-physics



Hi,

My name is Yann Pouillon. I'm coordinating the software activities of the European Theoretical Spectroscopy Facility (ETSF, http://etsf.eu/), a network of around 200 researchers all over Europe. The ETSF carries out state-of-the-art research on theoretical and computational methods for studying electronic and optical properties of molecules and materials. We work essentially at the nanometer scale, where most of the current challenges in our field are located. Our previous efforts were considering smaller systems, for which everybody could develop their own methods and software independently, but addressing the issues raised by the nanoscale requires much more cooperation and coordination.

In order to make it practically possible, we have created a common file format with its associated I/O library (based on NetCDF), rewritten some essential and common components as libraries, and made various utilities available to our community. At the same time, we have increased the quality of our software by following higher coding standard, and we try now to stick to the GNU Coding Standards as much as possible. We have also significantly improved our build systems, test suites and documentations. We have now reached a point where we have to deal with the distribution of our software, in a sensible and efficient manner.

Last year, when presenting the ETSF at the RMLL in Nantes, France, I had a discussion with Andreas Tille who proposed me to create a metapackage for Debian Science covering our field. After a lot of internal discussions and preparation, I'm glad to tell you that we have decided to provide and maintain two metapackages, "nanoscale-physics" and "nanoscale-physics-dev". During the last year, we have learned a lot about Debian and Debian Science, and checked that our software and development models were actually compatible with the philosophy of Debian and its requirements, and that the packaging of our codes would not put too much pressure on us. The latter was indeed the main concern of our developers.

The motivations behind having these two metapackages are the following:
We have categorized our software and devised a strategy to create the metapackages in the easiest possible way. We will start with all the tools and libraries our main codes depend on and, in a second stage, once all this first set of packages is available, we will start to package our main codes. A pilot project will be conducted with our most complex code, Abinit. We will thus progressively fill the metapackage. I'll be the contact person between Debian and the ETSF. I am participating to the creation of the packages and I'll maintain the metapackage.

Here is a tentative list of what could go in the "nanoscale-physics" metapackage:
To my knowledge, only Abinit and V_Sim are currently packaged within Debian. Our other dependencies, such as BLAS, LAPACK, ScaLAPACK, FFTW, GSL, MPICH, OpenMPI, ..., are already packaged within Debian.

We expect to start the creation of the metapackage in the second half of January 2011. But before, we would gladly appreciate your comments, questions and suggestions, in order to eventually make this project bear the most beautiful fruits.

Best regards,

Yann.

--
Free Software is good for you.

Reply to: