Network of SUN cpu's and GRASS

[jimg at lake.rutgers.edu]

It was mentioned that serving the binaries off of one machine onto clients
via NFS was somewhat system administrivia, ie mounting file systems and such.
Although we don't use this system with GRASS, we do with other GIS packages
and other utilities (ARC/INFO, ERDAS, OpenWindows, X11), not to mention data
areas, with no noticeable slowdown in performance.

I'd like to mention the use of automounting daemons for this whole shebang, 
which makes the mounting tasks somewhat invisible. Sun distributes its own
automounter (which can work fine at times), but we use the Berkley automounter
daemon (amd). The automounter mounts file systems when needed (such as cd'ing
into a directory or using mountable binaries), and unmounts them after a certain
time period when not in use (around five minutes). Figuring out the ins, outs,
ups, and downs of automounting can be a bit of a hassle, but if a geographer
like me can figure it out, then how hard can it be ?? (Where's that map again ?)

On that note, I once heard that the best thing that could happen to GIS would
be to get rid of Geographic, hence ridding the technology/discipline of 
numerous annoying pests.

Jim Gasprich
Cook College Remote Sensing Center
(which isn't to blame for these comments)