[gpaw-users] Summary of last weeks users and developer meeting

[Jens Jørgen Mortensen]

Hi!

Here is a very short summary of some of the discussions we had last week:

It was decided to only have one mail list for GPAW (and only one for 
ASE).  So, we'll disable the gpaw-developers and ase-developers ML's.

We will be supporting Python 2.6 for still some time because it's still 
used in some places.  It turned out that among the people present, only 
DTU-people are still using Python 2.6 on their supercomputers - very 
embarrassing!

We want to switch from GPAW's own vdW-DF implementation to an external 
dependency (libvdwxc), but only after we have figured out why the 
numbers differ.  Mikael, Ask and Johannes will look into this.

The PAW-correction to vdW-DF is not ready yet.  The suggested solution 
is to generate new PAW datasets that have more electrons in the 
pseudo-core density (similar to how PP's use a non-linear core 
correction).  Someone should document this ...

The new "augment_grids" keyword should be changed to a better name (I 
don't remember what we decided on).  The default value will be True if 
there are enough grid-points per CPU for this to be efficient.

We discussed how to help people who are forced to do their 
GPAW-development work on a MAC.  But no one wanted to document how they 
installed GPAW, so we are stuck with our broken documentation.

A small test among the people present showed that everyone had 
setuptools installed.  So, we might want to base our setup.py on this 
some time in the future (setuptools = modern distutils).

And finally: Thanks Lauri for arranging such a nice meeting!

Jens Jørgen