Hello, I am new to the list. Just to tell everyone who I am: I am a
research associate at the Institute for Fusion Studies, University of
Texas. I got my PhD in plasma physics 4 years ago, doing kinetic
simulation of plasmas (continuing to this day). I've been working
extensively with computers for the last 12 years or so. I've written a
largish (60K+ lines) plasma simulation code in MPPL (extended Fortran
variant), contributed greatly to a graphics package in C (20K lines), along
with other smaller projects.
I recently decided to take a hard look at NetCDF to serve my needs as data
file format for large plasma simulations.
I had previously been working on my own format, since I felt that was the
best for my purposes. This includes: (a) a lossy form of floating point
compression, (b) faster storage of floating points than XDR on non-IEEE
machines such as Crays (this turns out to be related to (a)), and (c) file
sequencing, where a "logical" file can be created in pieces, according to
some byte limit on each file. These features help a lot when dealing with
very large data sets (100's of MB to GB).
On the other hand, the NetCDF package seems to be well thought out, is of a
reasonable size, and the source code seems not too difficult to follow. So
I am considering how I might go about the task of enhancing NetCDF to do
the things I want. Which brings me to some questions that are more
"political" than technical:
o who are currently developing and maintaining NetCDF?
o are "outsiders" allowed to be part of the club?
o if so, does this include access to the central CVS or RCS repository?
o if not, (to either of the above), how are user-contributed enhancements
handled and how do the outside developers stay current with recent
versions of the code?
Your comment is greatly appreciated.
--
Maurice LeBrun mjl@xxxxxxxxxxxxxxxxxxxx
Institute for Fusion Studies, University of Texas at Austin
Emacs is a fine operating system, but I still prefer UNIX.
